KR20010081097A - Computerized tomography for non-destructive testing - Google Patents

Abstract

The present invention relates to a CT image reconstruction method wherein different values for at least one reconstruction parameter are used for at least two different portions of the image. The at least one reconstruction parameter preferably comprises a pixel size.

Description

Computed tomography and apparatus for nondestructive testing {COMPUTERIZED TOMOGRAPHY FOR NON-DESTRUCTIVE TESTING}

Various CT reconstruction techniques used when the purpose is to obtain an image slice of the entire body in situations where physical dissection is unnecessary are known in the art. The most important CT used is medical imaging when the patient's body cannot be dissected for obvious reasons. In a typical CT procedure a plurality of projections for an image slice are obtained and the slice is reconstructed from the obtained projection data. Projection data is achieved by trans-illumination of the slice using X-rays. In other medical imaging techniques, projection data is also obtained by radiation emission of radio-nucleotides from the body.

CT imaging has also been used to some extent industrially, such as nondestructive testing. However, the resolution of such a system is relatively low. Medical systems also have relatively low resolution, but they are generally optimized to minimize radiation exposure and provide suitable contrast for low contrast tissue imaging.

US Pat. Nos. 5,450,462, 5,379,333 and 5,400,378, incorporated herein by reference, describe a two-step CT imaging technique in which an initial scan without reconstruction is used to determine the appropriate radiation level in an actual imaging scan. The purpose of these patents is to minimize radiation levels while maintaining acceptable noise levels. Japanese Patent Publication Nos. 05305077, PCT Publication Nos. WO98 / 33361 and US Pat. Nos. 5,696,807, 5,228,070 and 5,485,494, which are incorporated herein by reference, describe other methods of determining to change the X-ray radiation intensity as desired. have. In particular, reconstructed slices of an object at the same or adjacent tomographic locations can be used to determine the desired radiation for a particular location.

High resolution CT imaging is achieved on small objects using micro focus X-rays. However, the number of pixels in the image is almost the same as in other CT imaging techniques. In general, the complexity of reconstruction increases with the square of the number of pixels, so reconstruction of a large high resolution image is impractical.

Two-dimensional X-ray photography for nondestructive testing is also known.

In general, an object prepared, including a cavity, is examined by dissecting the object and measuring on the dissected object. This inspection is important because the amount of raw material required for the manufacture of the object is defined by making a tolerance of the thickness of the part surrounding the cavity. Clearly, the test cannot be performed on all manufactured objects. When a CT scan examines an object with a cavity, two important issues are considered. First, a larger operating range is needed than before. Second, the reconstruction time is long and high resolution images are not obtained.

The present invention relates to a computerized tomography (CT) imaging technique, and more particularly, to high-resolution nondestructive examination CT imaging.

1 is a schematic diagram of a spherical object slice showing a deviation from a design specification.

2 is a flowchart of CT image reconstruction processing according to a preferred embodiment of the present invention.

3 is a flowchart of a reconstruction complexity reduction method according to a preferred embodiment of the present invention.

4 is a flowchart of a projection angle selection method according to a preferred embodiment of the present invention.

5A is a flowchart of a method of defining a multi-resolution grid according to a preferred embodiment of the present invention.

FIG. 5B is a schematic diagram of an image of an object slice, illustrating a multi-resolution grid defined by the method of FIG. 5A.

6A is a flowchart of a band constraint generation method according to a preferred embodiment of the present invention.

FIG. 6B is a schematic diagram of a banded image of an object slice, as determined by the method of FIG. 6A.

7 is a flowchart of an iterative reconstruction method according to a preferred embodiment of the present invention.

FIG. 8A is a flowchart of a method of setting a relaxation coefficient for iterative reconstruction of FIG. 7.

8B is a flowchart of a moment constraint generation method according to a preferred embodiment of the present invention.

9 is a schematic diagram of an imaging apparatus according to a preferred embodiment of the present invention.

10 is a schematic diagram of an embodiment of the present invention for quality control of extrusion molding.

11 is a schematic diagram of a conveyor belt embodiment of the present invention.

One feature of the preferred embodiment of the present invention is to provide a method of CT reconstruction that quickly reconstructs even in large image arrays. Alternatively or in addition, the method allows for reduced memory requirements for the reconstruction. In a preferred embodiment of the present invention the reconstruction method is used for industrial inspection of the manufactured object.

A feature of the preferred embodiment of the present invention provides a method of CT reconstruction using conventional knowledge about the object to be reconstructed. In a preferred embodiment of the present invention, the prior knowledge is provided from the design specification of the object. Alternatively or in addition, conventional knowledge is provided from a photograph of the object and / or its slices. Alternatively or in addition, knowledge is provided from the predicted manufacturing tolerances and / or predicted problem areas of the object. In a preferred embodiment of the present invention, knowledge is used to alter the reconstruction of an image of an object, including but not limited to reconstruction techniques, reconstruction constraints, reconstruction parameters, reconstruction resolution, data acquisition parameters, and / or post processing. Alternatively or in addition, to conventional knowledge of use, an estimated (approximate) image can be generated by rapid reconstruction of minimal projection data, for example by rear projection.

A feature of the preferred embodiment of the present invention provides a reliable method of reconstructing an image of an object even in the presence of inconsistencies in the acquired data. The mismatch is caused by noise. Alternatively or additionally, the discrepancy is due to the fact that almost all of the stagnation line is absorbed by the object of at least one projection point. Alternatively or additionally, the discrepancy is due to the partial or negligible absorption of radiation at the corners and / or other details of the object. In a preferred embodiment of the present invention said mismatch is resolved by improving the reconstruction of the problem area or discarding the bad data.

Aspects of certain preferred embodiments of the present invention relate to obtaining individualized data and / or providing reconstruction for different points of an image slice. Preferably, individual processing for each point of an image includes pixel size, reconstruction shape, reconstruction parameters, Local setting of one of an initial value, an expected deviation, and / or an accuracy level. In the preferred embodiment of the present invention, the individual processing is provided by capturing only the difference between the image slice inference and the actual photographing slice of the object. If there is little or no difference between the shot slice and the speculative image, it is preferable that the reconstruction is performed using a few resources (eg, memory and CPU cycles).

A "reconstruction parameter", referred to herein, is a parameter of the reconstruction method itself, and if the parameter is changed, modify the mathematical or procedural action of the reconstruction. For example, the number of repetitions, relaxation coefficients and stop conditions in ART are examples of reconstruction parameters. Examples of such parameters in backprojection are a weight (ray projection) to apply to the trace and a function to select which trace to apply. The filtering point of the "filtered" backprojection does not consider the reconstruction part, but rather is only a preprocessing step that does not affect the reconstruction, only the data used for the reconstruction.

In a preferred embodiment of the present invention, a portion of the speculative image slice is set to determine the deductive image value, and only other portions of the slice are reconstructed. Optionally or additionally, different portions of the image slice may be associated with an accuracy level for the deductive image density value of that portion. Optionally or additionally, different data acquisition methods may be applied to the projection or portion of the projection that includes data as the different image portions. Optionally or additionally, different reconstruction methods may be used for different portions of the image slice.

Aspects of certain preferred embodiments of the present invention involve reconstruction in a much smaller number than all pixels of an image slice. The image value of the cell may be inferred at a high accuracy level, and the pixels are preferably not reconstructed. Optionally or additionally, low quality reconstruction is used for pixels that have a low impact on the final use of the reconstruction image. An example of such use is corner measurement, and the sharpness of the corner does not affect the desired measurement of the wall thickness. Optionally or additionally, low quality reconstruction is used for low expected error levels in reproduction and / or image reconstruction.

Aspects of certain preferred embodiments of the present invention relate to a method of selecting a pixel on a reconstruction to be executed. In a preferred embodiment of the present invention the speculative image (or other representative value) of the object is used to determine the portion of the object or portions of the object that are likely to have speculative deviations. Depending on the characteristic thickness, location on the object, material and / or method of manufacture, the expected amount of deviation will be different for each part of the object. Optionally or additionally, the expected deviation is determined for the deviation allowed in the design at that location. In a preferred embodiment of the present invention the boundary of the object is determined. By being present inside the object (full pixel) or being outside the object (empty pixel), it is preferable that a pixel distant from the boundary is given a fixed value. Pixels that are not given a fixed value generally define a band around some or all object boundaries. However, in certain preferred embodiments of the present invention, the pixel to be reconstructed defines an out-of-band geometry, such as an island. It is desirable that only pixels within the band are reconstructed.

Aspects of certain preferred embodiments of the present invention relate to the reduction of reconstruction artifacts. In an embodiment of the present invention, the artifact is limited to only the portion of the image that is substantially reconstructed, and does not affect the portion of the image to which a fixed deductive value is assigned. Optionally or additionally, since the noise is limited only to the reconstructed region, the reconstructed image is relatively unaffected by the noise. Optionally or additionally, the fact that a clear portion of the image is known before the reconstruction is used to determine the appearance of artifacts and / or noise in the image, for example by comparing the reconstructed image with the speculative image. In a preferred embodiment of the present invention, spatial noise, distortion and / or artifact issues are detected in the part where the composition of the image is known. If artifacts are found, it is desirable that artifacts be extracted from the entire image. As an example, Compton scatter statistics are obtained by analyzing portions of the image that do not have expected deviations. Different statistics are collected for different parts of the image, and each reconstructed pixel is corrected using the Compton statistics of the unreconstructed pixel at hand.

Features of the embodiments of the present invention relate to a significant reduction in the computational complexity in reconstruction. As can be seen, the number of pixels in the band image is considerably less than the total number of pixels in the same image. The number of pixels to be reconstructed is reduced from n ² to kn, where k increases and k is determined by the complexity of the object. The complexity of the reconstruction made is nlogn. Additional techniques for further reducing the number of reconstructed pixels are also described herein. In a preferred embodiment of the invention, the reconstructed image is a binary image. Optionally or additionally, the reconstructed image is a discrete image with discrete image values. The number of discrete density values is equal to the number of materials in the imaged object, plus one for air. Optionally or additionally, a large number of discrete values may be used to correspond to the boundary area, for example, but the continuous range also functions. However, by limiting the number of density values, fast reconstruction is achieved in the preferred embodiment of the present invention.

A feature of the preferred embodiment of the present invention relates to a multi-stage image reconstruction method. In a preferred embodiment of the present invention, first, an estimated image is provided and then the estimated image (first image) is used to reconstruct at least an accurate second image in that predetermined portion. In the preferred embodiment of the present invention, an estimated image is provided based on the design specifications of the imaged object. Optionally or additionally, preferably using a low or low quality data and / or low quality reconstruction method, for example filtered rear projection, a small number of projections, low spatial resolution of the projection data, large (for reconstruction) Using pixel size and / or short data acquisition time is provided based on fast reconstruction of the image at the expense of quality. In certain preferred embodiments of the present invention, data is obtained at multiple projection angles, generating speculative data. Optionally or additionally, the angle used for reconstructing speculative data is determined by analysis of deductive knowledge, such as CAD design of the object. In certain preferred embodiments of the present invention the actual first image is not reconstructed but otherwise not provided. Rather, the second image is reconstructed using deductive information or a sinogram of the first image and / or partially reconstructed first image with selective lowness.

More accurate reconstruction is preferably based on speculative images (or deductive information). Pixels having values specifiable at high levels of accuracy for the speculative image are not reconstructed into the second image. In a preferred embodiment of the present invention, after more accurate reconstruction, the projection data used for reconstruction is cleaned, and a third, more accurate image is reconstructed. In a preferred embodiment of the present invention, the completion and / or reconstruction step is also performed using different completion techniques and / or parameters for at least one iteration as possible.

In a preferred embodiment of the present invention the final more accurate image is reconstructed so that the edge and / or other features of the image are extracted to perform the measurement. In a preferred embodiment of the present invention the edge is determined at the resolution of the subpixel using the moment. Optionally or additionally for edge determination, reconstructed images are used for other methods and / or uses.

In a preferred embodiment of the present invention the estimated reconstruction uses a filtered back projection method. Optionally or additionally, accurate reconstruction and / or more accurate reconstruction uses ART (Algebraic Reconstruction Technique) reconstruction methods and variations thereof. Preferably, the ART technique is applied repeatedly. Optionally or additionally, Bayesian or other repetition techniques may be applied that exemplify the maximum likelihood. Optionally or additionally, non-repeating techniques such as rear projection may be applied. In a preferred embodiment of the present invention the second, third and / or successive reconstructions do not all use the same reconstruction technique. Preferably, the result of the previous reconstruction is used as a starting point for the next reconstruction.

In the preferred embodiment of the present invention, the emission level used for each projection that acquires data second (and subsequent images if additional data is acquired for them) is determined based on the previous reconstruction. Thus, for example, an estimated image can be used to set the desired emission level so that the desired signal-to-noise level, operating range and / or sensitivity of the CT imager is achieved. Preferably, the emission level is selected to optimize a portion of the projection including the ray passing through the reconstructed pixel at the expense of the portion of the projection that does not include the pixel to be reconstructed. In a preferred embodiment of the invention, certain portions of the object are imaged at different radiation levels at different angles to finer the good image.

In a preferred embodiment of the invention the subsections of the imaged object are singled for specific reconstruction, for example at different emission levels and / or other resolutions of data acquisition and / or reconstruction.

A feature of the preferred embodiment of the present invention relates to setting an initial estimated density value for the pixel to be reconstructed. In a preferred embodiment of the present invention, the pixel to be reconstructed is characterized by being in a band around the object boundary. Preferably, each pixel is associated with a density value that responds to the distance from the object's nearest boundary and other properties. Optionally or additionally, the density value is a function of the band and / or other parameter of the object or part thereof, for example the width of its moment. The initial density value is used as an initial value in the reconstruction method of the ART type. In a preferred embodiment of the present invention the initial value is selected according to the moment of projection data.

A feature of the preferred embodiment of the present invention relates to forming an external reconstruction boundary “hull”. In a preferred embodiment of the present invention, the hull is used to fix the density value estimated in the reconstruction step. In a preferred embodiment of the present invention, the hull is formed to cover a subset of the image bound by the pixel to obtain a non-zero projection value. Therefore, the hull is approximated by a polygon formed by the outermost pair of radiating rays that yields a non-zero projection value. The initial and / or final density values for the pixel are set to zero. Optionally or additionally, during the repetition of the reconstruction, the estimated density value is made to minimize the previous repetition value and the hull value. In one example, if the pixel is hull or "zero" inside, the hull value is initially set to "0" if the pixel is outside hull. Optionally or additionally, the hull is formed to include two or more discrete values that calculate an estimated density value, for example, based on the identification of the plurality of materials and / or based on the estimated error. Optionally or additionally, for using a hull to fix a value, the hull is used to perform another mathematical operation of the pixel value, eg multiplication.

In certain preferred embodiments of the present invention, for example, if the photographed object falls into an X-ray attenuation flow, the pixels outside the envelope are clamped to non-zero values. The flow density is greater than the (X-ray) density of the object. Optionally, the flow density is less than the object density.

Certain preferred embodiments of the present invention relate to determining a constraint relaxation coefficient for a constraint according to an iterative reconstruction technique. In certain preferred embodiments of the present invention, the reconstruction techniques are other versions, such as ART-like reconstruction techniques, such as ART, MART, SART, ART-3, and / or algebraic reconstruction techniques. In certain preferred embodiments of the present invention, the reconstruction technique is applied iteratively to determine, depending on the previous iteration results as possible, different relaxation coefficients for each iteration. In certain preferred embodiments of the present invention, the relaxation coefficients in the ART-like reconstruction method are represented as a function of deviation of the mismatch values. The deviation is calculated for the allowed deviation gate, for example 3σ. In one example, the constraint is represented by the square of the deviation. Optionally or additionally, the constraint is limited at the maximum value. Optionally or additionally, the constraint is proportional to its maximum value.

Certain preferred embodiments of the present invention involve discarding some of the raw data obtained to improve the quality of the reconstruction. In certain preferred embodiments of the present invention, projections and / or projection portions (because their values are very low or noisy) that represent light rays passing through a large amount of material are ignored.

Certain preferred embodiments of the present invention relate to removing constraints in the ART-like reconstruction process. In certain preferred embodiments of the present invention, this constraint is suspected to be inadequate if the constraint causes deviation in the reconstructed image for projection data of a size larger than a predetermined value. Optionally or additionally, constraints are suspected according to their size for mean and / or other statistical matters. Optionally or additionally, deviations in constraint through several iterations may be considered. In a preferred embodiment of the present invention. Suspicious constraints are discarded only if they form a small percentage of constraints for a particular direction. Optionally or additionally, such discarding is based on the relative deviation of constraints from which projection data is generated from the same or near angle.

Certain preferred embodiments of the invention relate to using non-uniform pixels in reconstruction. In a preferred embodiment of the present invention, higher pixel resolution is used for pixels (e.g. for reconstruction or measurement) and / or pixels which are more important when the expected error or / and variation of local pixel values is greater. Optionally or additionally, pixel size is determined by the effect of lower pixel resolution on other reconstructed pixels, such as sharing the same radiation trace. Optionally or additionally, projection data may be obtained or stored at varying resolution levels. In one example, projection portions that pass through less important pixels in image use may be obtained or stored at lower spatial and / or density value resolution.

In certain preferred embodiments of the present invention, reconstruction data and / or projection data are stored using hierarchical data representations. Hierarchical representation maintains spatial organization of data. The data structure is a quad tree. Therefore, important data storage is only needed for "important pixels". If high resolution is provided to the band pixel, it is desirable. High resolution is preferably provided in pixels that are as close as possible to the speculative boundaries of the object being photographed, with respect to the expected error level and / or density level. In a preferred embodiment of the present invention, the local grid resolution can be changed during reconstruction. Optionally, a fixed (but varying spatially as possible) resolution is used. In a preferred embodiment of the present invention, the higher pixel resolution is approximately equal to the detector resolution. Alternatively, higher or lower resolutions can be used.

One aspect of certain preferred embodiments of the present invention relates to minimizing the number of projections required for reconstruction. In a preferred embodiment of the present invention, fewer than 360 projections are used for data reconstruction. It is preferred if fewer projections than 64 or 32 are used. In a preferred embodiment of the present invention, the projection is selected so that the projection is useful. In one example, the projection angle is discarded if a major portion of the obtained value is at or near the noise level, as a result of passing through the bulk attenuation material. Constraints in the projection angle can be applied to minimize the number of projections obtained. Alternatively or additionally, in a preferred embodiment of the present invention, constraints apply to minimize the number of projections participating in the reconstruction. In a preferred embodiment of the invention, the photographed object and / or its expected error is analyzed to determine a small number of projection angles suitable for reconstruction.

One aspect of certain preferred embodiments of the present invention relates to automatic constraints of an object taken tomographically in a coordinate system. In a preferred embodiment of the present invention, the object is placed on an imaging platform of any orientation and / or point. In a preferred embodiment of the present invention, coarse scan and reconstruction are used to guess the orientation. Optionally or additionally, the position can be inferred by comparing the size of the photographed object with the expected size and known fan beam geometry. Optionally or additionally, object identification for preferably selecting a speculative image, for example in a library, can be performed according to the results of the coarse scan. As will be appreciated the raw data (eg cosine curve) of the scan in many situations is suitable for this task without requiring an image to be reconstructed.

One aspect of certain preferred embodiments of the present invention relates to determining a threshold for binarizing an image (threshold for setting a discrete image value at two or more levels). In a preferred embodiment of the invention, the selected threshold is such that the zeroth order moment of the thresholded image matches some or all zeroth order moments of the original projection. The moment is found by binary searching the binary range between 0 and 1. Alternatively or additionally, higher order moments may be used for image binarization or other methods of image thresholding.

Therefore, according to a preferred embodiment of the present invention, an image reconstruction method for tomography includes the steps of providing an internal structure indication of a non-animal object to be photographed;

Selecting a projection angle for reconstruction according to the index;

Reconstructing an image from the data obtained at the selected projection angle, the method comprising:

Obtaining data at a plurality of projection angles,

Reconstructing a low quality image from the obtained data,

The indicator includes a low quality image. The data obtained at the selected projection angle preferably includes a subset of the data obtained at the plurality of projection angles. Optionally, each of the selected projection angles and the plurality of projection angles includes at least one projection angle that cannot be found at other projection angles.

In a preferred embodiment of the present invention, the data obtained with the plurality of projection angles includes a subset of all data obtained with the selected plurality of projection angles.

In a preferred embodiment of the present invention, selecting the projection angle includes selecting a minimum estimated number of projection angles that yields an appropriate reconstruction. Optionally or additionally, selecting the projection angle includes selecting a projection angle that responds to the object complexity. Optionally or additionally, selecting the projection angle includes selecting a projection angle responsive to the marker of heavy absorption at an index of the internal structure. Optionally or additionally, selecting the projection angle includes selecting a projection angle that corresponds to the low absorption index in the index of the internal structure. Optionally or additionally, selecting the projection angle includes selecting a projection angle that responds to a particular feature in the reconstructed image. Certain features may include at least one sharp edge. Alternatively, certain features include the boundary region of the object. The boundary area preferably includes a boundary between two materials containing the object.

In a preferred embodiment of the present invention, the step of selecting a projection angle includes selecting along the angle of incidence of the trace of the projection angle having the characteristics of the indicated internal structure.

According to a preferred embodiment of the present invention, the tomography image reconstruction method

Providing an internal structural indication of the non-animal object to be photographed;

Selecting at least one resolution in obtaining data according to the indicator; and

Reconstructing an image from the obtained data using the at least one resolution. At least one resolution includes at least two different resolutions for two different projection angles. Optionally or additionally, said at least one resolution comprises at least two different resolutions for the data obtained at a single projection angle. Optionally or additionally, the resolution includes spatial resolution. Optionally or additionally, the resolution includes grayscale resolution. Optionally or additionally, the resolution selection step includes selecting an estimated minimum resolution that can yield a suitable reconstruction. The complexity of the object preferably includes the complexity of the local object.

In a preferred embodiment of the present invention, selecting at least one resolution comprises selecting at least one resolution in response to the indication of heavy absorption in the indication of the internal structure. Optionally or additionally, selecting at least one resolution includes selecting at least one resolution in response to the indication of low absorption in the indication of the internal structure. Selectively or additionally selecting at least one resolution includes selecting at least one resolution in response to a particular feature of the reconstructed image. Preferably the particular feature comprises the reference feature used for the measurement in the image. Alternatively, certain features include border regions. Alternatively, certain features include self-measured features in the image.

In a preferred embodiment of the invention, the indication comprises an evaluation of the internal structure of the object. Preferably the evaluation includes the design specifications of the object. Alternatively, the evaluation includes at least a two dimensional representation of the object. Alternatively, the assessment includes a pre-reconstructed image of the object. Alternatively, the evaluation includes an image of a pre-imaged object of similar preparation.

In a preferred embodiment of the invention, the indication comprises a possible deviation from the desired internal structure. Optionally or additionally, reconstructing the image includes reconstructing only the portion of the object that responds to the indication.

Image reconstruction method according to a preferred embodiment of the invention,

Providing projection data of a non-animal object;

Reconstructing an image from the projection data,

Different reconstruction processes are applied to at least a portion of the image, and the different reconstruction processes include different reconstruction methods for the at least a portion.

Another image reconstruction method according to a preferred embodiment of the invention,

Providing projection data of a non-animal object;

Reconstructing an image from the projection data,

Different reconstruction processes are applied to at least a portion of the image, wherein the different reconstruction processes include using different values for at least one reconstruction parameter of the reconstruction method used for the at least a portion.

Image reconstruction method according to a preferred embodiment of the invention,

Providing projection data of a non-animal object;

Reconstructing an image from the projection data,

Different reconstruction processes are applied to at least a portion of the image, wherein the different reconstruction processes include a method of reconstructing the at least a portion at different spatial resolutions.

In a preferred embodiment of the invention, said at least part comprises at least two parts, each having different treatments applied to each other and to the third part of the image.

Image reconstruction method according to a preferred embodiment of the invention,

Providing projection data of a non-animal object;

Reconstructing an image from the projection data,

Different reconstruction processes are applied to at least a portion and at least a second portion of the image, and to the image at least three different processes, one for each portion and at least a different portion of the image.

In a preferred embodiment of the invention, different reconstruction processes include reconstructing at different resolutions.

In a preferred embodiment of the invention, the reconstruction method is

Providing an internal structure indication of the object,

The different reconstruction process is applied according to the instructions.

The image reconstruction method,

Providing projection data of a non-animal object;

Providing an indication of the internal structure of the object, and

Reconstructing an image from the projection data,

Different reconstruction processes are applied to at least a portion of an image, wherein the different reconstruction processes are applied according to the instructions.

Preferably, the indication comprises the step of reconstructing using a low quality image of the object. Preferably, the low quality image is reconstructed which is reconstructed using a reverse projection method. Alternatively, the low quality image is reconstructed using a Bayesian method. Alternatively, the low quality image is reconstructed using a Fourier transform method. Alternatively, the low quality image is reconstructed using the maximum likeness method. Alternatively, the low quality image is reconstructed which is reconstructed using the maximum entropy method.

In a preferred embodiment of the invention, the low quality image is reconstructed using a different resolution grid than that used in the image reconstruction step. Optionally or additionally, the image reconstruction step includes reconstruction using an algebraic reconstruction method. Alternatively, the image reconstruction step is reconstructed using the Bayesian method. Alternatively, the image reconstruction step includes reconstruction using a Fourier transform. Alternatively, the image reconstruction step includes reconstruction using a maximum likeness method. Alternatively, the image reconstruction step includes reconstruction using a maximum entropy method. Alternatively, the image reconstruction step includes reconstruction using a rear projection method.

In a preferred embodiment of the invention, the image reconstruction step includes reconstruction using a finer resolution than that used for the indication. Optionally or additionally, the spatial processing method includes setting an initial guess of the image according to the indication. Optionally or additionally, the step of providing instructions includes recovering the instructions generated upon imaging of a similar object. Optionally or additionally, the step of providing instructions includes providing a design specification of the object.

The internal structure instructions provided provide other examples or additional examples of manufacturing for the object. As another example or in addition, the special processing depends on an evaluation distance for at least one portion from at least one edge of the object, and the evaluation is based on an internal structure indication. As another example or in addition, the special processing depends on the expected reconstruction error for at least one portion, and the expected error is based on an internal structure indication. In another example or in addition, the special processing depends on the expected manufacturing error for at least one portion, and the expected manufacturing error is based on an internal structure indication. As another example or in addition, the special processing depends on the reliability of the internal structure of the object, and the reliability is based on the internal structure indication. In another example or in addition, at least one portion is at least one band surrounding the object. This band preferably includes an area that overlaps at least the outer edge of the object. As another example, the band includes an area surrounding the gap of the object.

In a preferred embodiment of the present invention, at least one portion substantially covers the characteristics of the object. This property preferably includes regions having a certain density.

In a preferred embodiment of the present invention, the special processing includes using to evaluate the selected pixel in at least one part without reconstructing the selected pixel. As another example or in addition, this special processing includes using to evaluate the selected pixel outside at least one portion without reconstructing the selected pixel.

In a preferred embodiment of the present invention, this special process involves a low level reconstruction outside of at least one portion. This lower reconstruction preferably includes a reconstruction with multiple errors. In another example, this lower reconstruction includes a lower resolution reconstruction of space.

In a preferred embodiment of the invention, the same preprocessing is applied to the inner and outer pixels for at least one part. As another example, different pixel processing is applied to the inner and outer pixels for at least one portion.

In a preferred embodiment of the invention, the preprocessing comprises filtering.

In a preferred embodiment of the present invention, the special treatment is at least partially in accordance with the level of detail required. As another example or in addition, the special treatment depends on the measurements performed on at least some of them. The object preferably conforms to at least one property of the composite material. At least one property includes the fiber direction of the material. In another example, the at least one property comprises the cell size of the material.

In addition, an image reconstruction method comprising: providing projection data of a non-animal object, providing an internal structure indication of an object; Reconstructing an image from the projection data,

The different reconstruction step includes only reconstruction of pixels from data in at least a portion of the image, the area being shaped in response to an internal structure indication. At least one particular region includes at least two discrete regions. As another example or in addition, the shape includes a band shape that encompasses at least one region of the pixel that is not reconstructed. As another example or in addition, this shape provides a preferred embodiment of the invention in which the boundary is determined from the boundary of the object indicated by the internal structural indication.

In addition, an image reconstruction method comprising: providing projection data of a non-animal object; Providing an internal structure indication of the object; A preferred embodiment of the present invention comprises reconstructing an image from projection data in response to at least one potential problem area in an internal structure indication. The at least one latent problem area preferably comprises an edge. As another example or in addition, the at least one potential problem area includes a suspected crack area. In another example or in addition, the at least one potential problem area includes a suspected void area. As another example or in addition, the reconstruction method includes changing the spatial resolution of the reconstruction, depending on the location of the at least one potential problem area. As another example or in addition, the reconstruction step includes changing the gray level resolution of the reconstruction, depending on the location of the at least one potential problem area. As another example or in addition, the method includes defining differently reconstructed regions in accordance with the at least one latent problem region. In another example or in addition, the method includes a method of determining a local confidence level according to the edge.

In addition, as a method of reconstructing the repeated image,

Providing projection data of the non-animal object to be photographed;

A first reconstruction step for the target object from the projection data;

Discarding at least some of the data in accordance with the first reconstruction; and

After the discarding step, there is provided a preferred embodiment of the present invention comprising repeating the first reconstruction step at least once.

In addition, as a repeated image reconstruction method,

Providing projection data of the non-animal object to be photographed;

Generating a reconstruction constraint;

First reconstructing the object from projection data in accordance with the reconstruction constraint;

Discarding at least some of the constraints in accordance with the first reconstruction; and

After the discarding step, there is provided a preferred embodiment of the present invention comprising repeating the first reconstruction step at least once.

In addition, as a repeated image reconstruction method,

Providing projection data of the non-animal object to be photographed;

Generating a reconstruction constraint;

First reconstructing the object from the projection data according to the reconstruction constraint;

Generating a relaxation coefficient in accordance with the first reconstruction; and

Using the relaxation coefficient, there is provided a preferred embodiment of the present invention comprising repeating the first reconstruction.

Further, providing projection data of a non-animal object to be photographed according to a preferred embodiment of the present invention, generating a reconstruction constraint, first reconstructing the object from the projection data according to the reconstruction constraint; And changing the first reconstruction value according to a dominant distribution function, and repeating the first reconstruction at least once after the changing step. The method preferably includes determining that the dominant distribution is zero outside a convex object defined by all traces of projection data having a zero projection value.

In a preferred embodiment of the invention, the method comprises processing the data before repeating the first reconstruction. In another example or in addition, the first reconstruction includes a plurality of repetitions. In another example or in addition, the repeated reconstruction according to the first reconstruction includes applying a different pretreatment to the first reconstruction of the object. In another example or in addition, the repeated reconstruction according to the first reconstruction includes applying different preprocessing to the projection data. In another example or in addition, the repeated reconstruction includes applying different preprocessing to the first projection data of the object, in accordance with the internal structural indicators of the object. In another example or in addition, the method includes applying different preprocessing to the projection data according to the internal structural index of the object.

Further, according to a preferred embodiment of the present invention, obtaining a projection data set, reconstructing a first image from the projection data using a first reconstruction method, and analyzing the image to a part of the image. A method is provided for obtaining an image of an object comprising the steps of determining a special processing method and reconstructing a second image of the object using the analysis, wherein the second reconstruction is different from the first reconstruction. Way. Preferably, the method includes obtaining data for the second reconstruction, according to the analysis. In another example or in addition, the data is obtained according to the desired analysis of the second image.

In a preferred embodiment of the present invention, the method comprises, in accordance with the analysis, varying the amount of ion radiation used to obtain the data. In another example or in addition, the method includes changing, according to the analysis, the wavelength of the ion radiation used for obtaining the data. It is preferable that the said ion radiation contains X-rays. In another example or in addition, the ion radiation includes gamma rays.

In a preferred embodiment of the present invention, the data is obtained using non-ionized electron radiation.

In a preferred embodiment of the invention, the method comprises changing at least one parameter of the detection circuit, in accordance with the analysis. The at least one parameter is preferably a gain.

In a preferred embodiment of the invention, the method comprises selecting at least one of the detection systems from the plurality of available elements, in accordance with the analysis, preferably the element is a detector, another example or further As such, the element comprises a filter. In another example or in addition, the element comprises an aimer.

In a preferred embodiment of the present invention, the second reconstruction method includes an algebraic reconstruction method. The algebraic reconstruction method preferably includes an ART-like method.

In a preferred embodiment of the present invention, the second reconstruction method includes a Bayesian reconstruction method. In another example, the second reconstruction method includes a Fourier transform reconstruction method. In another example, the second reconstruction method includes a maximum likelihood reconstruction method. In another example, the second reconstruction method includes a maximum entropy reconstruction method.

In a preferred embodiment of the present invention, the second reconstruction method uses a different resolution grid than the first reconstruction method. In another example or in addition, the first reconstruction method includes a rear projection reconstruction method. In another example, the first reconstruction method includes a Bayesian method. In another example, the first reconstruction method includes a maximum likeness method. In another example, the first reconstruction method includes a maximum entropy method. In another example, the first reconstruction method includes a Fourier transform method.

In a preferred embodiment of the present invention, the first reconstruction fails to achieve sufficient convergence for at least a portion of the image. The failure is preferably determined from the error level. In another example or in addition, the failure is determined from an unexpected value for the reconstructed pixel value.

Further comprising: obtaining a projection data set;

Reconstructing a first image from the projection data using a first reconstruction;

Analyzing the image to determine special processing for the image portion;

The data acquisition arrangement comprises a device selected from an available set of functionally identical devices, wherein the projection data acquisition stage acquires data for a second reconstruction using a data acquisition arrangement different from the first arrangement used;

A preferred embodiment of the present invention includes reconstructing a second image of the object. In another example or embodiment, the different configuration uses a different optical detector than the first configuration. In another example or embodiment, the different configuration uses a different filter than the first configuration. In another example or embodiment, the different configuration uses a different detector circuit than the first configuration.

In a preferred embodiment of the present invention, the method comprises post-processing the reconstructed image using an image processing method. The image processing method is preferably adapted to improve feature measurement in the reconstructed image.

In a preferred embodiment of the invention, the object is a manufacturing object. This object is preferably a cast object. As another example, this object is an outgoing object.

In addition, as a CT imaging device,

x-ray radiation source;

A data acquisition system comprising at least one set of functionally identical devices, comprising: a data acquisition system for acquiring attenuation data in accordance with the x-ray density of an object disposed in the system;

Preferred embodiment of the present invention comprising a controller for selectively selecting a particular one of the devices to be used in the data acquisition system according to its own analysis of data acquired through the system comprising a different one of the devices. To provide.

In addition, as an algebraic CT image reconstruction method,

Providing projection data;

Generating a constraint on the moment of data; and

It provides a preferred embodiment of the present invention comprising the step of reconstructing the image from the data using the moment constraint. In another example, the moment includes a primary moment. In another example, the moment includes a secondary moment. In another example, the moment includes a higher moment than the second moment.

Also, as a relaxation coefficient generation for an iterative algebraic reconstruction method,

A preferred embodiment of the present invention includes providing a previous image reconstructed by setting a constraint value during a constant repetition, and setting a relaxation coefficient for continuous repetition according to a change value between the constraint value and a value in the image. to provide. The relaxation coefficient setting step

And comparing the constraint value to a threshold and setting a relaxation coefficient in accordance with the comparison. The threshold is preferably a statistical principle function of the variation.

In addition, a CT image reconstruction method including providing projection data of a non-animal object from a plurality of projection angles and back projecting the data into a data structure displaying a variable resolution grid is provided according to an embodiment of the present invention. do. The data structure preferably includes a hierarchical data structure.

Furthermore, according to a preferred embodiment of the present invention,

Feeders of one or more non-animal objects to be photographed, and

An industrial inspection system is provided that is mechanically connected to the feeder and includes a CT imager for imaging the one or more objects. The source preferably comprises a conveyor belt for carrying a plurality of objects. Optionally or additionally, the object is assembled from subcomponents. Optionally, the object is cast. Optionally, the object is rolled. Optionally, the object is injection molded.

In a preferred embodiment of the invention, the feeder comprises a take-up device.

In a preferred embodiment of the invention, the object is mechanized.

In a preferred embodiment of the invention, the feeder comprises an extrusion machine for extruding the object in the form of a continuous profile. The extruder preferably comprises an extrusion nozzle. Optionally or additionally, the extruder includes a shaper.

In a preferred embodiment of the present invention, the object consists essentially of one material. Optionally, the object consists of a composite material. Optionally, the object consists of two or more materials.

In a preferred embodiment of the present invention, the CT imager is tuned to photograph an object according to the object supply of the feeder. Optionally or additionally, the CT camera captures the object using a spatial imaging method. In a preferred embodiment of the present invention, the CT imager uses the method described herein.

In addition, in a preferred embodiment of the present invention, a CT camera for the CT industrial imaging field,

A detector;

an x-ray source;

An image area;

And a manipulator for supporting and transporting an object to be projected to or from the image area. Optionally, the manipulator comprises a winch.

Further, in a preferred embodiment of the present invention, the quality assurance manufacturing method is

Preparing an object;

Projecting the object with a CT camera;

Measuring a feature relating to the image to detect a deviation of an internal structure of the object from a design specification;

Rejecting the object if the object does not match the design specification. Preferably, the object is cast. Optionally, the object is extruded. Optionally, the object is mold molded.

In a preferred embodiment of the invention, the method

Analyzing the image to detect at least one material defect;

Rejecting the object in response to the detected defect. Preferably, said at least one defect comprises a bubble. Alternatively or additionally, the at least one defect comprises an empty space. Alternatively or additionally, the at least one defect comprises a change in density. Alternatively or additionally, the at least one defect comprises a crack. Alternatively or additionally, the at least one defect comprises a crust.

According to a preferred embodiment of the present invention,

Obtaining photographic data from a plurality of projection angles of the non-animal object;

First reconstructing an image from the photographing data using a tomographic reconstruction method, and

A tomographic reconstruction method is provided that includes introducing a discontinuity into the image to convert an image value of the reconstructed image into a restricted setting of an allowed image value, wherein the discontinuity maintains at least one image characteristic. The at least one image characteristic includes a moment of the image. It is preferable that the moment is a primary moment. Optionally, the moment is a second or higher moment.

In a preferred embodiment of the present invention, the constrained setpoint includes only two values. Alternatively or additionally, the method includes a second reconstruction of the image iteratively after the discontinuity introduction step. Alternatively or additionally, the constrained setpoint depends on the number of values and a value relative to the expected density value of the image.

Furthermore, according to a preferred embodiment of the present invention,

Providing a non-animal object having at least one subcomponent in known geometry;

Obtaining photographing data of the object from a plurality of projection angles; and

A tomographic reconstruction method is provided that includes reconstructing an image of the object using the at least one subcomponent of known geometry. Video reconstruction method,

Generating a low quality image reconstruction;

Identifying sub-components of the image; and

Using the identification, in principle, further comprising reconstructing the image. Identifying the subcomponent preferably comprises identifying a location of the subcomponent. Optionally or additionally, the identifying subcomponent comprises identifying the direction of the subcomponent.

In a preferred embodiment of the present invention, the further reconstruction step includes fixing a value for a pixel of the image, wherein the pixel corresponds to a portion of at least one subcomponent, and the fixing step is the fixing step. According to the geometry.

In a preferred embodiment of the invention, the object comprises essentially at least one subcomponent. Alternatively or additionally, the at least one subcomponent comprises two subcomponents in known geometry.

The invention will become more apparent from the following detailed description of the preferred embodiment and the accompanying drawings.

1 is a schematic diagram of a spherical object 20 slice showing a deviation from a design specification. Object 20 includes a wall 22 that surrounds cavity 24. The dashed line 28 shows the design specification of the object 20, and the line 26 shows the deviation during manufacture. In the casting object, there is no method of measuring the cavity 24 directly without slicing while breaking the open object 20. Much of the cost of the casting object is a result of the uncertainty in the inner wall thickness. Since these walls require a minimum thickness, casting objects are often manufactured using wall thicknesses that are thicker than necessary to compensate for possible errors. It should be noted that the thicknesses and variations involved are often small. The deviation is less than one tenth of a millimeter, and for a given object, its cross-sectional area may comprise several centimeters of material.

In a preferred embodiment of the present invention, an object such as object 20 is inspected using X-ray computed tomography (CT), which is a non-destructive testing method. Preferably, a dedicated CT inspection system is provided. 2 is a flowchart of CT image reconstruction processing according to a preferred embodiment of the present invention. First, the method will be described in more general terms with reference to FIGS. 3 to 8.

First, a speculative image of an imaged slice of an object 20 is obtained (30). In a preferred embodiment of the present invention, the speculative image is generated by low quality imaging of the object 20. In one embodiment, a small number, such as 16 projections, is used to reconstruct the image slice. In one embodiment, a filtered backprojection reconstruction method is used. When the reverse projection reconstruction technique is used, since the image is blurred by reverse projection, an edge detection method is performed on the image.

Alternatively or additionally, the object 20 is a known object made, for example, in accordance with design specifications. In a preferred embodiment of the present invention, design specifications and / or CAD files are analyzed to provide the estimated image. Alternatively or additionally, low quality data acquisition and / or reconstruction is used to determine the similarity between the design specification and the photographed object to determine whether the design specification can be used for coarse guessing of the object cross section. In a preferred embodiment of the present invention, reconstruction is not necessary because the sinogram of the obtained data is analyzed to determine similarity in the design specification and / or CAD file.

In a preferred embodiment of the present invention, low quality data acquisition and / or reconstruction is used to identify which object is being photographed. In a preferred embodiment of the present invention, this identification can be performed by matching the obtained cosine curve to the virtual cosine curve of the design specification without reconstructing the speculative image.

Referring to Figure 2, as will be apparent from the description of certain preferred embodiments of the present invention, the more known similarities between "guessing" images and actual image slices, the lower the complexity of image reconstruction that can be achieved in certain cases.

Then, in accordance with a preferred embodiment of the present invention, it is preferred that the speculative image is analyzed 32 to determine the likelihood and / or limit of applying a method of lower complexity. In a preferred embodiment of the invention, the analysis includes edge detection of the object boundary.

Additional projection data can be obtained (34). Projection angles and / or other parameters for the obtained projections are determined in accordance with analysis 32 and / or data currently available. In one embodiment, the radiation trace through a large amount of material is attenuated at the system noise level so that the projection angle is data that is not available or available. In yet another embodiment, the sensitivity of the detector arrangement may be preset at the expected radiation transmission level. In another embodiment, the source radiation level can be set according to the expected attenuation.

After data is obtained, the data is reanalyzed to determine parameters that reduce the reconstruction complexity. This reduction is achieved by lowering the number of pixels to be reconstructed and / or their resolution and / or the accuracy of the reconstruction.

The obtained projection data is then reconstructed as reconstruction "A" 38. In a preferred embodiment of the present invention, iterative reconstruction is used, preferably an ART or ART-like reconstruction method is used. The speculative image is used as a starting point for repetition. Optionally, starting points are generated in response to speculative images and / or deductive knowledge. Discrete reconstruction methods are used. Optionally, a continuous reconstruction method is used. The description of ART is a description, which is incorporated herein by reference, in GT Herman's "Image Reconstruction from Projections", 1980's "The Fundamentals of Computerized Tomography" or Y. Censor's "Finite Series-expansion Reconstruction Methods (preceedings of the IEEE, Vol 71, No. 3 Mar. 1993, pp. 409-419).

In a preferred embodiment of the present invention, after the reconstruction "A" is completed, an optional second reconstruction step is performed. The second reconstruction step includes discarding data 40 and / or reconstruction constraints that reveal the problem and performing reconstruction “B” 42. Preferably, the second reconstruction step is also ART reconstruction. In certain preferred embodiments of the present invention, one or two reconstructions may use a non-ART or even non-repetitive reconstruction technique. However, the advantage of using ART-like techniques is that they are suitable for parallel processing. In addition, certain ART-like techniques are free to select a reconstruction start point.

Alternatively or additionally, other preprocessing operations, particularly cleaning operations, may be performed on the partially reconstructed image and / or projection data, in order to discard data and / or reconstruction constraints. In one embodiment a smoothing filter may be employed. The preprocessing is preferably in accordance with the partially reconstructed image-for example the determination of the problem area in the aspect. For example, smoothing does not apply in this area. Alternatively or additionally, pretreatment is in accordance with previously provided indicators for imaging metabolites. For example, smoothing does not apply to data on complex or reconstructed parts of an object.

When the image slice is reconstructed, it is preferable that the edges in the image slice are determined to reconstruct the contours of the object 20 and 40 slices. The reconstructed image is then measured and used to determine the deviation from the design specification and the reconstructed object.

3 is a flow chart of a method of reducing reconstruction complexity (36 in FIG. 2). According to a preferred embodiment of the present invention, although the steps of FIG. 3 are shown in a particular order, it will be apparent from the following description that these steps may be performed in any order or in two steps combined in a single step. In addition, some or all of these steps are performed with the analysis of the speculative image, for example, if the object specification has been entered into the system, even before the speculative image is obtained 30. Each step of FIG. 3 can be applied individually. In particular, in some embodiments of the present invention, only one or even none may be applied in the step of FIG. 3.

The steps in FIG. 3 are as follows:

(a) selecting 50 a constrained number of projection angles;

(b) using a multi-resolution grid to set non-uniform pixel size (52);

(c) a band image generation step 54, which is a subset of the completed image slice;

(d) setting 56 reconfiguration parameters for various regions of the image slice, eg, in accordance with a trade-off between image quality and reconstruction complexity or according to expected local error.

4 is a flowchart of a method of selecting a projection angle (50 in FIG. 3) according to a preferred embodiment of the present invention. Selecting a projection angle (i) reduces the number of projections used for recovery, preferably reduces memory and / or CPU requirements, and (ii) uses more preferred projections to reduce errors and ensure fast convergence. It serves two purposes. Preferably, some or all of the projection data is obtained only after selecting the projection angle. Additionally or alternatively, the projection data can be obtained initially, and only data that matches the desired projection angle is used for recovery.

In step 60, sinograms, evaluation images, and / or prior knowledge of formulated data, etc., are analyzed to determine suitable and / or unsuitable angles. In a preferred embodiment of the present invention, the angle at which significant traces are completely attenuated by the object 20 is considered unsuitable. Preferably, the significant traces are those that have passed important pixels (eg, based on user indications and / or automated analysis) and / or those that have passed many pixels without many relevant traces. Additionally or alternatively, the angle is selected such that the dynamic range of values expected in one projection is minimized and consistent with system limits. Alternatively or additionally, the angle is selected such that the traces do not pass through an area that is difficult to recover. Alternatively or additionally, the angle is chosen such that the portion of the image of interest is actually illuminated by independent traces in the range from which the appropriate image can be recovered.

In a preferred embodiment of the present invention, the preset number of contributions is 64 or 32. The possible projection is preferably 62 according to step 60.

Next, the 32 or 64 best angle is preferably 64 is selected. The preferred selection range coincides with the distribution function of the particular angle beyond the range of possible projection angles of that selected angle. Additionally or alternatively, the projection angles are ranges that do not overlap together.

Additionally or alternatively, the number of projections may be determined depending on the image of the objects being made. One way to select multiple projection angles is to change the number of projections and / or the particular projection angle to select an angle that yields satisfactory results and reduces computation time while simulating a recovery. Optionally or additionally, the projection can be selected by analyzing the geometric shape of the slice depicted by the object and / or image.

In a preferred embodiment of the invention, an object depicted as an image is placed in the system in a random direction, and a low resolution image and / or sinogram is used to evaluate the direction, for example the selection of the projection angle. And / or facilitate traces of previous shape, such as CAD design.

In a preferred embodiment of the present invention, when a particular projection angle is desired, a plurality of projections can be obtained at approximate angles, and at the highest and / or most accurate projection angles used for recovery. Therefore, the accuracy of the angle needs to be low.

In a preferred embodiment of the invention, the slices depicted in the image are selected to match the slices in the design specification. Thus, minimal deviation is expected. In addition, any determined deviation is more important if the deviation is indicated on the design drawing of the slice. In a preferred embodiment of the present invention, the report of the results depicting the object visually includes a visual, statistical and location specific analysis of the deviation found.

Other steps 52, 54, and 56 of FIG. 3 involve identifying and selectively handling specific portions of the image slice. In a preferred embodiment of the invention, the selection of the projection angle is made in response to this step. For example, the data acquisition dynamic range described above can only be determined for traces that have passed through an important area. Additionally or alternatively, the projection angle is selected based on how the data of this particular part is obtained, and the whole is unnecessary for the image.

5A is a flow chart of a method of forming the various maritime grids 52 of FIG. 3 in accordance with a preferred embodiment of the present invention. 5B is a schematic diagram of an image of an object slice 70 showing various resolution grids 72 in accordance with the method of FIG. 5A.

In a preferred embodiment of the present invention, smaller pixels are assigned to a portion of the image slice that requires high resolution of the image. Preferably, the pixel size is determined by the range of expected local error and / or density values. In a preferred embodiment of the invention, the object 20 is imaged to detect deviations from known designs. Thus, high resolution of the image is typically required at or near the highlight of the object slice 70. Alternatively, if the design of the object 20 is unknown, certain heuristics may be applied to the image based on the task variable. For example, in some object slices, one cannot expect to be an island smaller than any size. Therefore, based on the coarse recovery at the resolution of the size, it is possible to determine all the image areas in which pixels are not empty. Note that in a typical object observation application, there are two or three materials in the object 20, such as metal, air and possible spacers. Preferably, pixels with relatively specified values are large, while pixels where values are expected to change are small. Typically, pixels at the boundary of the object are expected to change, but pixels away from the boundary are expected to remain constant.

5A illustrates a method of providing an object slice boundary small pixel. In addition, such a small pixel is preferably set in advance to a guess density value based on the guess image. Optionally or additionally, the pixel value may take into account the expected deviation in manufacturing the object. Minimum and maximum pixel sizes are set (80). The speculative image slice is divided into super-sized pixels, and an initial density value is set for each large pixel (82). For each pixel, if the pixel is not the minimum allowable size, and if its density (normalized) is not 0 or 1, the pixel is preferably divided into four or nine small pixels as much as possible (84). If the pixel size is larger than the predetermined threshold size, each pixel is assigned 86 a calculated "real" density value. The amount of "black" (attenuation) in the area covered by the pixel divided by the pixel size is the density value.

If the pixel size is less than or equal to T1, the virtual density value is calculated. The virtual density is preferably the density value of the divided pixel. Optionally, other density functions may be used, for example the pixel value will be a function of distance to the boundary. T1 is preferably the minimum pixel size. Additionally or alternatively, T1 may vary spatially, so that, for example, it will be a function of distance to the boundary (or band—see FIG. 6A).

The process is repeated until all pixel values have a value of zero or one or are at a minimum size. In a preferred embodiment of the invention, all pixels having a density value of 0 or 1 and not of minimum size do not change during recovery. The projection data is a sensitivity corrected to ignore the pixel, for example to reduce attenuation value of the projection data.

In a preferred embodiment of the present invention, T1 and the minimum and maximum pixel sizes are the same over the entire image. Alternatively or additionally, different values are used for different parts of the image.

In FIG. 5B, grid 72 shows the results of providing the method of FIG. 5A (only a portion of the image slice). Pixel 74 is essentially the largest size, pixel 76 is the medium size, and pixel 78 is the minimum size. In the method of Figure 5A, the pixels on the boundary are automatically determined to be of minimum size. However, in the preferred embodiment of the present invention, the other pixels can be set to medium size. Also, in some instances and circumstances, pixels on the boundary are allocated to a size larger than the minimum size in case a deteriorated recovery is needed at that point.

5A and other methods may be used to define the pixel size in the preferred embodiment of the present invention. As an example, the pixel size may be formed to monotonically increase the function of the distance of the pixel starting at a distance of one standard deviation (expected or acceptable error) from one boundary. Alternatively or additionally, the method of FIG. 5A can be modified such that one pixel is divided into small pixels in response to the required quality of recovery, and the pixel is divided into pixels whose recovery is significantly affected.

The pixels are square pixels, and in each step of the method of FIG. 5, the pixels are divided into four equal parts. Optionally or additionally, non-square pixels such as octagons or other polygons may be used. Optionally or additionally, five or nine methods of segmentation and / or asymmetric segmentation may be used. Optionally or additionally, the pixels do not fit into the regular grid. Optionally or additionally, the pixels have non-uniform resolution, for example, may have a higher value along one axis of the object than the other axis. Optionally or additionally, different pixel grids may be used for different portions of the image, and the grids are preferably selected to match the image reconstruction conditions. Optionally or additionally, the pixels may overlap. Optionally or additionally the pixels are stored as pixel center values, the pixel values being determined using interpolation between pixel center values or between derived values.

6A is a flow diagram of a method of defining a band (step 54 of FIG. 3), in accordance with a preferred embodiment of the present invention. FIG. 6B schematically illustrates the band point 100 of an image slice, according to the method of FIG. 6A. As mentioned above, various deviations in the object manufacturing process occur at the boundary of the object. Assuming that the deviation is limited to a certain amount, some pixels of the image slice are not expected to contain any material (empty pixels). Optionally or additionally, other portions of the image may be expected to include full pixels. In FIG. 6B, two bands 104 and 102 are defined near the portion indicated with reference numeral 100. The density of the pixels (boundary pixels are known to have an intermediate density) is suspected in this band. In the center region 103 (indicated by the diagonal lines), all pixels are expected to be cold. These pixels (areas 103, 105) are not modified because they are preferably fixed during reconstruction.

Non-zero density values of the pixel should be allowed outside of the band. However, in many cases, only zero values will be outside of the band. Optionally or additionally, the tolerance value is continuous and not necessarily limited to zero or one, outside and inside the band. If the full range of deviations is found in some, the reconstruction quality at that point will not be high. However, this object will probably be defective, and the exact level of the defect is not very important. Therefore, in some embodiments of the invention, the large pixels will be tolerant of the outer range of the band.

Referring to FIG. 6A, the first step is to determine an "error point", the deviation may be expected or generated 106. As can be appreciated, different degrees of variation (or density values) in different portions of the photographed object are expected or allowed. Optionally or additionally, certain portions of the object will be predefined to be easily troubled and will require higher quality reconstruction and wider tolerance ranges. In some cases, some pixels may be determined to have a known density. In one embodiment, the photographed object will include a jig having exactly known dimensions. The jig's pixels are preferably identified or removed from consideration. Alternatively or additionally, certain pixels may be reconstructed even if they are outside of their defined band. In one embodiment, in casting, if an air bubble (pore) of a predetermined size is expected, a wall having a thickness no greater than twice the predetermined size may be reconstructed to ensure that it does not contain bubbles. Such reconstruction will be less limited with respect to density values and attenuation.

In a preferred embodiment of the present invention, an expert system is provided that can analyze a slice of an object (inferred image), a manufacturing method, and a previously detected deviation to produce a map of a problematic object area. Optionally or additionally, the map may be provided by a human expert.

Optionally or additionally, the band size or accuracy level is generated by simulated imaging of the object. In a preferred embodiment of the present invention, the system includes a simulator for determining, for example, the result of the shooting angle and the pixel size of the deviation detection, which can test different errors and shooting conditions.

In a preferred embodiment of the invention, the reconstruction and / or small pixel size will be limited to the portion of the object to be measured. Projection data for radiation trajectories that do not pass through the region is preferably ignored.

In a preferred embodiment of the present invention, when a multi-resolution grid is applied in FIGS. 5A and 5B, the bands may be extended to align with pixel boundaries and / or pixels of at least minimum size.

In a preferred embodiment of the present invention, the spatial resolution of the reconstruction varies with the determination of the problem area. Optionally or additionally, by improving detection sensitivity and / or gain for traces through possible problem areas, the resolution of the gray level will be higher for problem areas.

In step 108, at each pixel, a level of consecutive accuracy is assigned to the value, based on possible speculative images. In general, as shown in FIG. 6B, pixels closer to the object are more likely to be colder than pixels further away. This type of dependency may change over the entire image, taking the local error condition as far as possible. Additionally, for example, the inner pixel of the object and the outer pixel of the object may be different from each other while reflecting a material that is more likely to escape inside the edge on the substrate than the rest of the material is present, and in a preferred embodiment of the present invention, the edge and Linear dependence on distance is used for distance with an accuracy level value. Optionally, squared or exponential dependencies are used in the relationship between distance and accuracy. Optionally or additionally, multiple (possibly reduced) bands may be defined, each having its own level of accuracy and / or zero order approximation. In a preferred embodiment of the present invention, if the accuracy level is used to constrain the reconstruction process, it is more difficult to fill cold pixels of low accuracy.

In certain preferred embodiments of the present invention, the accuracy level is used only for the first configuration, for example to set the minimum density value. Optionally or additionally, the level of accuracy changes during reconstruction. Optionally or additionally, the same level of accuracy is used throughout the reconstruction.

In the preferred embodiment of the present invention, since the accuracy level is used as the probability field, the calculated density of the pixel is the product of the reconstructed density value and the accuracy value. Such values can be applied to the speculative image between reconstruction iterations. Optionally or additionally, the level of accuracy is a value setting for each pixel representing the probability of the density value within a predetermined range. Therefore, some pixels may have a high accuracy of "1", while others may have a high accuracy of "0.5" (eg, other materials) and may be "0". In some cases, since the accuracy level includes a muti-moda distribution function, the pixel may have a value between 0.1 and 0.2 or between 0.7 and 0.8 above any intermediate value.

In a preferred embodiment of the present invention, different radiation levels are determined with different projections, for example to maximize detector sensitivity. In a preferred embodiment of the present invention, only radiation level optimization and / or other imaging parameter optimization allows for imaging of pixels within the band.

7 is a flowchart of an iterative reconstruction method, in accordance with a preferred embodiment of the present invention. In a preferred embodiment of the invention, certain projection data is first removed from consideration. Preferably, a hull is created. The envelope is preferably defined as a convex geometric object surrounding the metabolite and there may be a pixelless exterior of the object. Projection data from traces that do not intersect the exterior are preferably removed from consideration. The envelope is defined for the object containing the band. Optionally or additionally, the envelope is preferably defined by analyzing the direct projection data, preferably only on the object itself, excluding the band. Optionally or additionally, the envelope is defined to include a different width of the band than is used for reconstruction. In a preferred embodiment of the present invention, the envelope is used to define a critical dominant distribution function for the reconstructed density. During the reconstruction, at each iteration, preferably before each iteration, each pixel is preferably assigned a minimum value between the reconstructed density value and the envelope value of the pixel. Optionally or additionally, each pixel value may be clamped in the expected value range (eg, between 0 and 1). In a preferred embodiment of the present invention, the envelope may include a plurality of values for different pixels, so that one pixel may be limited to a maximum value of "0.5" and another pixel to "0.9" (clamped). ). The number of clamped values will depend on the number of materials and / or the level of accuracy. In a preferred embodiment of the present invention, pixels located at (or near) the boundary of the reconstructed image are not clamped.

In a preferred embodiment of the present invention, under consideration, an initial value is set 112 for each pixel. Some of these values are set using the method of FIG. 5A. In a preferred embodiment of the present invention, the initial value is determined such that at least a predetermined moment of projection is maintained. Optionally or additionally, all projection moments are maintained. The retained moment is preferably the zeroth order moment. However, primary, secondary or higher moments can also be maintained.

In a preferred embodiment of the present invention, an iterative reconstruction technique such as ART is used. In a preferred embodiment of the present invention, each radiation trace is modified for beam hardening using, for example, a lookup table (and / or previous guessing of the image). Preferably, the lookup table is calculated. Optionally or additionally, the lookup table is derived experimentally during the measurement of the imaging device. In the ART reconstruction technique, the relaxation coefficient is preferably used to help converge the repetition into the final image. In a preferred embodiment of the present invention, the relaxation rate reset step proceeds after every one or more iterations (114).

Alternatively or additionally, the intermediate reconstruction technique may be binarized during the iteration step (or may be converted into phase with two or more discrete values). Preferably, as described below, such binarization preferably uses a threshold to preserve the moment of the reconstructed image and / or projection data.

In a preferred embodiment of the present invention, repeating the reconstruction technique is repeated until the escape condition is satisfied. In a preferred embodiment of the present invention, the escape state is a state in which the perfect error between the reconstructed image (or the binarized version of the image) and the projection data is smaller than the predetermined error level. Optionally or additionally, the exit state includes reconfiguration time and / or computer operation limitations. Optionally or additionally, the escape state determines whether the amount of change in error over the iteration is less than a predetermined amount or range. The predetermined amount described above may be, for example, a function of counting the reconstruction step performed, the complexity of the image and / or the reconstruction parameter.

In a preferred embodiment of the present invention, the local (or global) image complexity is to measure the number of edges crossed by tracking. Another measuring method is to measure the density of the corners in the area. Another possible measurement method is to measure the average shape size in the form of an edge or a line. Another possible measurement method is to measure the average deviation of local image measurements. The image complexity is measured by combining any one or more of the above-described measurement methods and / or image complexity measurement methods well known in the field of image processing.

In a preferred embodiment of the invention, the reconstruction of the first portion of the image comprises an error and an unacceptable second portion in the first portion of the image that is acceptable, for example, while the second portion of the image continues to be reconstructed. Can be stopped or stopped depending on the error in. The image may preferably be divided such that the trace does not apply to the other part but crosses only in one part.

Optionally or additionally, different reconstruction techniques can be applied to different portions of the image. Optionally or additionally, different techniques may be applied to different image slits and / or different iterations.

Although described with reference to FIG. 2, the reconstruction includes two or more sets of iterations, with problematic constraints in between (40 of FIG. 2). In a preferred embodiment of the present invention, after a certain iteration (eg 20 times), certain constraints are identified as the problem is revealed. In a preferred embodiment of the present invention, the constraint is identified if the constraint deviates from the reconstructed image by a threshold outlier. The threshold is a function of the number of repetitions and / or the communication theory of mean and / or deviation.

In a preferred embodiment of the invention, statistical theory calculates the deviation for this and / or other parts of the image or for the entire image. Optionally or additionally, the statistics belong to this and / or projection values for the entire image or other portions of the image. Optionally or additionally, the statistics belong to image values reconstructed for the partial or full image. Optionally or additionally, this identification is a function of the statistical theory of the deviations to the constraints over the various iterations to identify problem constraints during a given iteration. For example, constraints in the deviation cycle between extremes can be problematic constraints. In a preferred embodiment of the present invention, the constraint is removed if the problem constraint does not include a clear portion of the entire path in the given projection direction. Clear parts may be defined as 30%, 20% or 10% or more of the data. Optionally or additionally, if enough data is available at a close projection angle, the constraint is removed even though it forms 30% of the data. Optionally or additionally, if the problem constraint is temporarily removed, the removal is switched if this removal does not help convergence.

Optionally or additionally, problem constraints are identified based on the presence of an unexpected pattern in the partially reconstructed image. In one embodiment, the island may be of an unexpected nature. In another example, a spike, e.g., a pixel with one value that is surrounded by a pixel with zero value, may be considered to be an error. In another example, sharp corners are often unrealistic. Optionally or additionally, the unexpected pattern is determined based on manufacturing techniques and / or other deductive knowledge such as CAD manufacturing.

In a preferred embodiment of the present invention, the pharmaceutical system is adjusted by removing the highest constraint (the one with the highest degree of restriction). Optionally or additionally, the greater the projection data, the more the constraints. Optionally or additionally, at least the reconstruction for that area (pharmaceutical) is repeated, for example if the area is identified as including a corner, then the spatial reconstruction parameter for the corner is used so that there is no restriction to be removed.

In the case of certain problem constraints, detector noise, detector blurring, indeterminate detector response, source noise, dispersion and / or diffraction, thickness resolution limitations, very high or very low absorption (thick plates or corners), (e.g. Due to inhomogeneous material, poor measurement, poor object positioning and / or orientation, vibration and / or movement.

Optionally or additionally, to remove the constraints, other thinning operations such as image flattening, pixel value averaging, edge detection and enhancement and thresholding may be performed between the image reconstruction steps to assist in convergence.

In a preferred embodiment of the invention, the pixel size does not change during the overall iteration. Optionally or additionally, pixel size and / or band width may vary. In one embodiment, the pixel and / or its neighbors are added if the value in the pixel is surrounded by a pixel that reaches 1 (or 0) (within given precision) or the density value of the pixel is 1 (or 0). In repetition, it is removed from consideration. Optionally or additionally, the band can be widened if several nonzero pixels intrude the edge of the band. Optionally or additionally, the pixel may be reduced in size if the value does not converge to zero or one.

FIG. 8A is a flowchart of a method of setting a relaxation coefficient during the iterative reconstruction (step 116) of FIG. 7. Relaxation coefficients are often used in ART techniques to assist in convergence. Relaxation to get satisfactory convergence It is well known in the art that = 0.05 should be used. In a preferred embodiment of the invention, different Values are used for different constraints. This different value relates to the size of the deviation between the constraint and the reconstruction value of the projection. This relaxation coefficient should be updated and set at least once in a repetition step.

First, the deviation of the reconstructed value is determined (step 120). These deviations are preferably compared with each other or with previous values (step 122). It is desirable that a new relaxation coefficient be determined (step 124). Preferably this new relaxation coefficient is limited in the range of 0-1, preferably 0.05-1. In a preferred embodiment of the invention, it is only a problem if the deviation is present outside a certain gate, for example at a 3σ gate (around the expected value). In a preferred embodiment of the invention, the relaxation coefficients increase the linear function of the deviations. Secondary relationships are optionally used for linear relationships. Optionally, an power relation or exponential relation function is used. Different relationships between the deviation and the relaxation coefficient can be defined for different parts of the image. In a preferred embodiment of the present invention, if the relaxation coefficient cannot be improved by changing the relaxation coefficient for a specific restriction, the relaxation coefficient may be kept at a previous value or may be set to a predetermined value, for example, 0.05.

Referring to FIG. 8B, in a preferred embodiment of the present invention, the moment of limitation may be defined, for example, all regions (masses) of the image may be defined. In a preferred embodiment of the present invention, zero order moment is used. Alternatively or additionally, higher moments are used, such as primary or secondary moments. In a preferred embodiment of the invention, this moment is the average of the moments for some projection directions, especially when unbalanced beams are used or noise occurs. Thus, the projection moment is calculated (step 126), and a constraint extracted from this moment is generated (step 128), which limits the reconstructed image as a whole rather than as individual pixels. Alternatively or additionally, moment constraints or other multicell constraints may be defined for a portion of the image.

Often, reconstruction is completed at this point. However, in certain preferred embodiments of the present invention, further reconstruction steps may be performed next, as described above with reference to the cleaning operation between reconstruction steps, for example, to clean an image.

In a preferred embodiment of the present invention, the image is checked to determine if it matches the projection data after reconstruction. The reconstructed image is projected at the projection angle, so the degree of inconsistency between the projection data and the generated projection is important. In a preferred embodiment of the present invention, the degree of mismatch is used to inform the end user of the expected error level in the image and / or measurement. Optionally or additionally, if the degree of mismatch is higher than a predetermined value, the reconstructed image is discarded and additional data is required. In a preferred embodiment of the present invention, only a predetermined portion of the image is checked. Alternatively or additionally, different degrees of inconsistency are allowed for different parts of the image.

In a preferred embodiment of the present invention, the image is binarized or thresholded to include values of two or more discrete sets, so that the image contains only discretely set values, each of which preferably corresponds to a material in the object. In a preferred embodiment of the present invention, the threshold is determined so that the zeroth order (and / or higher order) moment of the image does not change by the thresholding action (within a predetermined level of precision). In a preferred embodiment of the present invention, a suitable threshold is found by performing a binary search in the threshold range between 0 + ε and 1-ε.

In a preferred embodiment of the invention, the edge is detected in the resulting image. These edges are used to perform measurement and / or other inspection tasks. Preferred forms of edge detection are described in J.C., which is a description herein incorporated by reference. It is described in The Image Processing Handbook (p. 674, 2nd Ed., CRC Press, 1994) by Russ. Optionally or additionally, the image can be analyzed using another tool, such as a pattern recognition tool. Optionally or additionally, imaging is used for non-testing tasks, such as enhancement or genealogy purposes.

In a preferred embodiment of the present invention, one or more of the following measurement techniques are applied to the reconstructed image.

(a) Mechanical Properties. For example, the distance between the entini and the feature

In general, this measurement is necessary for edge determination.

(b) geometric shapes, such as circles or tubes

For example, such measurements are also necessary for edge detection and for high quality / high resolution reconstruction for a slice of a portion.

(c) non-uniformity of density

In a preferred embodiment of the invention, the reconstructed image is used as a starting point for further reconstruction, and portions of the image suspected of containing bubbles or nonuniform densities are allowed to change this density value during reconstruction.

In a preferred embodiment of the present invention, depending on the measurement required, various reconstruction parameters may be changed or relaxed, allowing for faster and / or more effective reconstruction. In one embodiment, lower pixel resolution may be used at least at the corners if the features of the corners are not needed in the final image. Optionally or additionally, there may be interaction between the measurement and imaging device. For example, the imaging device may be measured so that the x-ray beam is aimed at the highest point at the angle of projection that results in a trace passing through the main pixel at the expense of other pixels in the image slice as much as possible. In some cases, the aimer must be used to be perpendicular to the x-ray beam for this purpose, so that the high quality part of the aimer interacts with the beam of the required part.

In one form of measurement, the next question is: That is, a simple yes / no answer to "Is this object within the tolerance of the design?" By tailoring the reconstruction quality / precision so that the reconstruction error and the actual error sums are below the allowable error for the image according to the whole or the main part of the image, this question can often be answered using reconstruction below the maximum quality. If the answer is yes, it does not matter how much deviation is in the manufactured object. If the answer to the question is "no" with a low enough probability, then the subject should be left to the laboratory where more accurate tests will be applied.

Optionally or additionally, the reconstructed image is analyzed to identify problem areas such as, for example, discrepancies or unexpected absorption. That part requires more attention. In a preferred embodiment of the present invention, particular attention is paid to repeating the reconstruction of the part, so that new data is available at better statistics at higher resolutions as possible. In the example of bubbles, large bubbles can be observed by using high resolution. Small bubbles that cannot be observed affect the average density and should be measured. Other problematic areas include areas in which two materials or two sources, such as plastic, and additives such as color or plasticizers or two separate aluminum billets are mixed. In some cases, an intermediate density value may be acceptable for certain parts of the object (for reconstruction), but not for other parts. Optionally, the median represents a potential problem area.

In a preferred embodiment of the present invention, additional reconstruction steps may be required for some portions of the image. In the case of corners of complex parts of the image, data must be obtained at additional and / or different projection angles. Optionally or additionally, very dense or very dense parts (eg corners) will be outside in the dynamic range of the imaging system. In a preferred embodiment of the invention, this portion is taken again with higher (or lower) intensity x-rays so that data is obtained within a dynamic range.

In a preferred embodiment of the present invention, the interface area between the two materials is analyzed to detect the boundary portion. Optionally or additionally, the interface region is analyzed to determine the presence of crust to form or separate between the two materials. As noted, the boundary thickness is much smaller than the typical pixel size used for image reconstruction. In a preferred embodiment of the present invention, a small pixel size is used for the border region. Optionally or additionally, at least one projection angle is selected such that the x-ray beam is substantially parallel to at least some boundary region. In a preferred embodiment of the present invention, the enhanced data acquisition resolution is only used along the boundary portion. The boundary portion preferably comprises a plurality of boundary regions selected as statistical samples of the major overall boundary regions. By analyzing these regions, it is possible to determine the probability of defects along the boundary and / or their average thickness. Optionally or additionally, such statistical analysis approaches may be introduced throughout the entire image to determine other types of discrepancies or find defects between the reconstructed image and the object design.

9 is a schematic diagram of a photographing apparatus 200 according to a preferred embodiment of the present invention. As a CT imager, the device 200 uses an x-ray tube (or other source) and some type of detector 216. In a preferred embodiment of the present invention, the device 200 uses very high aimed beams to achieve high spatial resolution. Optionally or additionally, the scintillation detector does not provide the required resolution. In the embodiment of FIG. 9, the device 200 includes three x-ray sights and an optical sight for the CCD based detector 216. Device 200 includes a computer 218 that controls device 200 or reconstructs an image from data obtained therefrom.

The first x-ray aimer 204 is disposed at the outlet of the tube 202. The second x-ray aimer 206 is disposed between the tube 202 and the imaged object 208. The third x-ray collimator 210 is preferably placed behind the object 208 to remove dispersion and diffraction at the object 208.

The patterned x-ray beam is preferably detected using a screen-camera combination. Screen 212 is preferably irradiated by a beam. Light exiting the screen is preferably aimed by an optical sight 214 (x-ray radiation absorption), which is detected by the linear CCD camera 216. In a preferred embodiment of the invention, the screen is read by a laser beam. Alternatively or additionally, the screen is a fluorescent screen, which emits light in a pattern of x-ray beams for a short time after irradiation. Optionally or additionally, the screen acts as a scintillator crystal and the object re-examines each data line obtained.

In a preferred embodiment of the invention, the CCD camera has a high spatial resolution, for example 20 micrometers and a number of elements, for example 10,000. In a preferred embodiment of the present invention, a TDI type CCD camera is used to compensate for any movement in the image and to be adapted to the camera's transverse movement and to improve camera light sensitivity. The cause of the movement is vibration. Optionally or additionally, a two-dimensional CCD camera is used. In certain embodiments of the present invention, camera 216 is smaller in screen than screen 212, so camera 216 must be scanned mechanically or selectively across the surface of screen 212. The scan should only be done on one axis. Optionally or additionally, the scan is two dimensional. Optionally or additionally, camera 216 includes a zoom lens to control over the entire size of the object and to adjust the data acquisition resolution.

In a preferred embodiment of the present invention, object 208 is moved to position using conveyor belts, cranes, robotic arms, and / or other technically known actuators. The movement of the object must be adapted to shooting so that the shooting is taken as soon as the object is positioned. This can be accomplished using central control of both the feeder and the object. Optionally, the camera's sensor activates the camera as the object passes the sensor. In a preferred embodiment of the invention, the projection angle is reached by moving the detector and / or x-ray tube. Alternatively or additionally, the object is moved and / or rotated.

In a preferred embodiment of the present invention, the device 200 is checked and / or calibrated using x-rays. Preferably a illusion is used. Alternatively or additionally, device 200 may be tested using an optical test pattern instead of screen 212. Alternatively or in addition, a light source may be used instead of the object 208.

In a preferred embodiment of the present invention, the device 200 uses a parallel beam of x-ray radiation. Alternatively or additionally, the device 200 uses a fan-beam configuration. Alternatively or additionally, the device 200 utilizes a scanning pencil beam.

As can be appreciated, the calibration and / or operation of device 200 may interact with the selected reconfiguration method. In one example, if only a small portion of the object 218 is simulated, the collimator can be closed, so that only a thin x-ray beam is produced and less diffraction and dispersion are produced. Conversely, if the collimation of the beam is limited in quality, the size of the smallest pixel will be larger. Alternatively, smaller pixel sizes can be used to compensate.

In a preferred embodiment of the present invention, the apparatus 200 includes multiple detection path elements, each optimized for a specific data acquisition condition. In a preferred embodiment of the present invention, the path element is automatically selected by the device in response to the data acquisition request. In one example, the device 200 includes a plurality of collimators, such as, for example, high resolution and low resolution collimators. As another example, the device 200 may comprise a plurality of optical detectors of different spatial, temporary and / or gray level resolutions and / or different optical sensitivity of different response sensitivity, gain control, sensitivity range, spectral response and / or other detection parameters. And a detector. In another embodiment, the device 200 includes a multi-space or x-ray wavelength filter (not shown). In another embodiment, apparatus 200 includes multiple signal processing circuitry (eg, embedded in computer 218). In another embodiment, the device 200 includes a multiple scintillation or solid state radiation detector 216.

10 is a schematic diagram of an embodiment of the present invention for quality control of extrusion molding. In a preferred embodiment of the present invention, the extruded object 222 is extruded by a nozzle 224 connected to the material source 220. In a preferred embodiment of the present invention, since the CT imager is positioned as indicated by reference numeral 228, the cross section of the object 222 is imaged. Optionally or additionally, the CT imager is located at the location indicated by reference numeral 226 so that both object 222 and nozzle 224 are imaged simultaneously. Therefore, problems (eg, bubbles) and / or nozzle wear of the extrusion process itself can be detected.

In a preferred embodiment of the present invention, object 222 is in continuous movement, and a helical CT reconstruction technique is used. In this reconstruction method, virtual projection data is generated by an interpolator between a close projection angle and / or a position of the object 222. At position 226, the nozzle does not move, so helical scanning does not apply. Optionally or additionally, in the device of FIG. 9, while the object 208 is located inside the device 200, helical scanning is used to image the object. The helical scanning is useful for more effective volumetric imaging. Optionally or additionally, helical scanning and / or reconstruction is used to cancel the request to stop moving the object 208 and / or components therein (because the object is brought inside) before imaging. Optionally or additionally, for helical scanning, defects in the object 222 have axial resources to consider, and therefore the movement of the object 222 should be considered to be negligible during reconstruction. In particular, inconsistencies during reconstruction may indicate defects among them.

In a preferred embodiment of the present invention, the object 222 image is reconstructed substantially in real time, such as 10, 5, 1 or 0.1 or less. In addition, the reconstructed number of pixels is at least 1,000,000, 20,000,000 or 100,000,000 pixels. This type of resolution and reconstruction time can also be used to photograph moving objects, such as a moving clock. It can be seen that the use of the method described herein and the main part of the photographed object have a known geometry that is closely or accurately known, and the actual processing requirements are preferably reduced compared to conventional methods.

In a preferred embodiment of the present invention, CT cameras are used to photograph objects made of one or more metals, glass, plastics, rubber, concrete, wood, composites and / or other industrial materials. Optionally or additionally, the imager is used to image an object comprising a plurality of materials, such as aluminum, for example made of plastic spacers. In a preferred embodiment of the present invention, since the photographed object is a non-animal system, the human body of life and death is excluded. In a preferred embodiment of the present invention, the composite material is photographed in a way that reveals information about the composite material. In one embodiment, high spatial resolution may be provided along the direction of the material particles (eg pixel size). Optionally or additionally, in order to better observe the particles of the individual component materials, a micro size is provided at least in part of the object. In a preferred embodiment of the invention, in particular the radiation in which the polymeric material is used for imaging may be used to generate radiation treatments, for example crosslinking in the polymeric material.

Objects photographed by the CT cameras described herein may or may not be cast, compressed, micro-extruded, roller-milling, laminating, machining and / or using connectors. In fact, it can be produced using any known manufacturing method including, but not limited to, assembling. In particular, certain preferred embodiments of the present invention are suitable for continuous processing such as compression molding as well as discontinuous processing such as casting. In a preferred embodiment of the present invention, the casting object can be photographed during casting. In addition, for certain casting processes, single slide imaging may be sufficient to demonstrate the speciman.

The photographed object has a cross section at 10-100 cm or larger than 20, 60 or 100 cm. The resolution is preferably 0.001 to 0.1 mm, such as 50, 20, 7 or even 1 μm. At a given address, the object is too large to be photographed at once, and overlapping portions of the object may be photographed. In one embodiment, the object is rotated (rather than a detector and / or x-ray source). If possible, knowledge of the internal source of the object is used to infer the orientation of the object. Optionally or additionally, a rotation sensor on the rotary actuator used to rotate the object is used to set each position of the object. In some embodiments of the present invention, slices of a plurality of axes of an object are obtained using a multi-axis swapped array. Alternatively or additionally, a plurality of axis alternating overlapping detector arrays can be used to receive radiation from an object, and can be used to read a phosphor display that receives radiation from the object.

Alternatively or additionally, to image relatively large objects, small objects may be imaged, including, for example, minerals, mineral samples, gems (either polished or unrefined) and / or artificial gems. Such analysis is preferably used to find hidden flaws, bubbles and / or structures.

11 is a schematic illustration of a conveyor belt embodiment of the present invention. In this embodiment, the conveyor 240 carries the object 242 to be imaged to the CT imager 230. In a preferred embodiment of the invention, all objects on the conveyor are imaged. Alternatively or in addition, for example, only samples of the object responding to the reconstruction time constraints of the camera 230 are imaged. Alternatively or additionally, only a sample of the produced object is delivered along conveyor 240 to be imaged.

In a preferred embodiment of the present invention, a fast image of the object 242 may be made utilizing multiple radiation sources (not shown) and / or multiple detectors and / or multiple detector rows (ROW). Alternatively or additionally, the fast image utilizes an electron beam CT imager in which a ring shaped target is scanned using the electron beam to create a fast moving source. For example, in FIG. 11, an electron beam from a source at position 232 is directed to a target 234 (shown in dashed line) located along a semicircular ring. One such image may also be used in the embodiment of FIG. 10.

In a preferred embodiment of the invention, the operation of the CT camera can be controlled by the user. Examples of inputs that may be provided by the user include the number of projections, the quality of the image required, the portion of the image that is of interest, the estimated image, the estimated image correction, the identification of the problem portion of the object, the reconstructed image correction and / or the CT imager. Include answers to questions raised by Such a question may be raised by the CT camera, for example, if the camera stops or becomes unable to reconstruct the image. The answer to this question includes, for example, the choice between two alternative reconstructions or resets of specific reconstruction parameters.

In a preferred embodiment of the present invention, each object to be imaged may utilize different reconstruction and / or acquisition parameters. In a preferred embodiment of the present invention, the parameters are selected based on the simulation run on the object design. Simulation can be used, for example, to determine the parameters that lead to the fastest reconstruction, the highest probability of detection error, the guarantee of convergence and / or the shortest data acquisition time.

Alternatively or additionally, these parameters can be determined by providing heuristic teaching to the image. One example of such heuristic teaching is to set the width of the band lower in the linear region of the image than in the curved region of the image.

In the same case, the imaged object may be blurred as a result of the moving artifact. In a preferred embodiment of the present invention, such moving artifacts can be corrected by using the knowledge that the imaged object is a hard object and assuming the movement vector and the movement of the projection data to calculate the motion vector in the calculation. Perhaps the motion vector is determined by analyzing the correspondence between the settings of the projection data from the image and / or other angles. Alternatively or additionally, the direction of movement can be known, for example, as a result of the movement of the CT camera and / or during the imaging. Alternatively or additionally, the motion vector can be determined by including data projected from the opposite side of the object. Alternatively or additionally, the motion vector can be determined utilizing the estimated image and / or other known knowledge. Differently or additionally, the motion generated by the vibrations can be corrected to measure the vibrations and estimate the motion blur generated thereby. In some cases, this modification may take into account different resonance characteristics of different parts of the object. In general, it is important that this analysis is performed on a single object and then applied to all created objects. In the form of nondestructive testing, the object is analyzed to determine the vibration mode of the underlying elements of the object. One or more images are obtained by subjecting the object to shocks and / or vibrations to test the object specimen. Manufacturing differences (deviations) between objects often cause vibrations in the resonant properties of parts of the object. Therefore, it is preferable that the obtained image is used as a "correct" object as the speculative image for the test image.

The present invention has been primarily described using 2D X-ray CT imaging. However, it can be understood that the above reconstruction technique is used for other imaging techniques. In one embodiment, the above reconstruction technique can be used to generate a cine image of the operation of the object. In this context, one form of deductive knowledge in a manufacturing object is the presence of a given object (subcomponent). If the layout of each of these objects is known and not in their exact location, the object can usually be identified from a low resolution reconstructed image. Once the object is identified, it should be possible to more accurately specify the position and / or orientation in the image, for example by using pattern matching to create as strong a constraint on the image as possible. For manufactured objects, component objects (usually subcomponents of manufactured objects) are often known, even if they are not in their exact location. In addition, the operating machine, while the (predetermined) subcomponent is in motion, its exact shape may be known (eg, the previous image). Optionally or additionally, the gap in the amount of motion of the elements is known (hence defining the band size as possible). Optionally or additionally, imaging is synchronized with the movement of the subcomponents. In some cases, the periodic or other normality of the subcomponents is determined by Fourier analysis, which analyzes a series of images to determine the frequency of wandering to certain parts of the object, or by analyzing the vibration vectors of the objects. Optionally or additionally, for cine imaging of the operating machine, the stressed object is imaged to track its failure mode. In all of the above cases, it is possible to limit the number of pixels that actually need to be reconstructed.

Optionally or additionally, to obtain a cine image, a 3D image of the object is obtained. One method of obtaining 3D images is the helical scanning described above. Optionally or additionally, multi-slice images may be obtained, if possible, by using a multi-row detector and / or by moving the imager relative to the object and / or each other and / or by using a cone beam. In an improved embodiment of the invention, the multi-resolution grid defined for the 3D image is 3D. For example, the band portion in one slice may depend on the object profile of a nearby slice. Such dependencies can be used when determining expected error and bandwidth in 2D imaging.

Optionally or additionally, a 3D image of the object can be obtained by photographing the object at an angle that is not all arranged on a single plane. In an extreme example, the object can be photographed in three vertical and intermediate directions, providing as much hemispherical coverage of the object as possible.

In a preferred embodiment of the invention, the imaged object can be prepared by imaging a liquid or gas in contrast media, for example by impregnation or immersion. This may be useful when imaging very transparent or porous structures. Alternatively or additionally, only one of the plurality of source materials mixed together to form the final material is tagged using contrast media. Therefore, the diffusion of the source material in the mixing can be determined.

Alternatively or alternatively, several embodiments described above for imaging using gamma radiation may be used for imaging x-ray CT. Alternatively or additionally, emission imaging is applicable for transmission imaging by using radiation source material instead of contrast media. Alternatively or additionally, SPECT imaging can be applied using the reconstruction technique. Alternatively or additionally, multiwavelength and / or combined gamma radiation and x-ray radiation are applicable. Alternatively or additionally, optical tomography can be performed. However, in general, when the photographed object is radiation, it is not commercially available. Therefore, this type of imaging is preserved for testing purposes and not for manufacturing purposes.

Certain embodiments of the present invention are particularly suitable for industrial imaging due to the limitations of type, requirements, allowed radiant energy, type of movement, predictability of structure and / or prior knowledge of the field. However, the above features and preferred embodiments of the invention are applicable to tomographic reconstruction, such as types of electronic tomography and / or medical imaging.

It will be appreciated that the CT image reconstruction method described above can be varied in many ways, including changing the order of multiple steps in which some steps are performed online and some steps are performed offline. Moreover, variously distributed and / or centralized configurations may be used, preferably, to realize the invention using a variety of software tools. Moreover, the various features of the various adaptations of both the method and the apparatus have been described so far. In particular, all the features shown in a particular embodiment are essential in all good, similar embodiments of the invention. Combinations of the above are also contemplated to be within the scope of preferred embodiments of the present invention. Also within the scope of the present invention are computer readable media having recorded thereon software for carrying out some or all of the preferred embodiments of the present invention. In addition, it should be understood that several embodiments are disclosed by only the method or by the device only. In addition, the scope of the present invention encompasses hardware and / or software that has been adopted and / or designed and / or programmed for practicing the embodiments of the method form. Moreover, the scope of the present invention includes methods of using, configuring, adjusting and / or maintaining the devices described herein. As used in the claims, the terms "comprise", "comprise", "have (or have)" and the like mean "including but not limited to".

Claims (203)

As a reconstruction method for tomography, Providing an internal structural indication of the non-animal object to be photographed; Selecting a projection angle for reconstruction in response to the indication; and And reconstructing an image from the data acquired at the selected projection angle. The method of claim 1, Acquiring data at a plurality of projection angles; Reconstructing a low quality image from the acquired data, wherein the indication comprises the low quality image. 3. The reconstruction method according to claim 2, wherein the data acquired with the selected projection intensity includes a subset of the data obtained with the plurality of projection angles. 3. The method of claim 2, wherein each of the selected projection angle and the plurality of projection angles includes at least one projection angle not found with each other. 3. The reconstruction method according to claim 2, wherein the data acquired at the plurality of projection angles includes a subset of all data acquired at the selected plurality of projection angles. 6. The reconstruction step according to any one of claims 1 to 5, wherein the projection angle selection step includes an estimated minimum number of projection angle selection steps that will yield a suitable reconstruction. 7. The reconstruction step according to any one of claims 1 to 6, wherein the projection angle selection step includes a projection angle selection step according to the complexity of the object. 8. The reconstruction step according to any one of claims 1 to 7, wherein the projection angle selection step includes a projection angle selection step that follows a serious absorption instruction in the indication of the internal structure. 8. A method according to any one of the preceding claims, wherein the projection angle selection step comprises a projection angle selection step that follows a low absorption indication in the indication of the internal structure. 10. A method according to any one of the preceding claims, wherein the step of selecting a projection angle comprises the step of selecting a projection angle following a special feature in the reconstructed image. The method of claim 10, wherein the special feature comprises at least one sharp corner. The method of claim 10, wherein the special feature comprises a boundary region of the object. The method of claim 12, wherein the boundary region comprises a boundary between two materials comprising the object. The projection angle selection step according to any one of claims 1 to 8, wherein the projection angle selection step selects a projection angle according to the incident angle of the trace of the projection angle having the characteristics of the indicated internal structure. Reconstitution. As a method of reconstructing an image for tomography, Providing an internal structural indication of the non-animal object to be photographed; Selecting at least one resolution in data acquisition according to the instructions; and And reconstructing an image from data acquired using the at least one resolution. 16. The method of claim 15, wherein said at least one resolution comprises at least two different resolutions for two different projection angles. 17. The method of claim 15 or 16, wherein the at least one resolution comprises at least two different resolutions for data acquired at a single projection angle. 18. The method of any one of claims 15 to 17, wherein the resolution comprises a spatial resolution. 19. The method of any one of claims 15-18, wherein said resolution comprises grayscale resolution. 20. The method of any one of claims 15-19, wherein selecting at least one resolution comprises selecting an estimated minimum resolution that can yield a suitable reconstruction. 21. The method of any one of claims 15 to 20, wherein the step of selecting at least one resolution comprises at least one resolution depending on the complexity of the object. The method of claim 21, wherein the complexity of the object includes the complexity of a local object. 23. The method of any one of claims 15 to 22, wherein selecting at least one resolution comprises selecting at least one resolution in accordance with a severe absorption indication in the indication of the internal structure. 23. The method of any one of claims 15 to 22, wherein selecting at least one resolution comprises selecting at least one resolution in accordance with a low absorption indication in the indication of the internal structure. 25. The method of any one of claims 15 to 24, wherein selecting at least one resolution comprises selecting at least one resolution in accordance with a special feature of the reconstructed image. 27. The method of claim 25, wherein the special feature comprises a reference feature used for measurement in an image. 27. The method of claim 25, wherein said special feature comprises a border region. 27. The method of claim 25, wherein the special feature comprises the feature itself measured in the image. 29. The method of any one of the preceding claims, wherein the indication comprises a guess of the internal structure of the object. 30. The method of claim 29, wherein said inference comprises a design specification of the object. 30. The method of claim 29, wherein said inference comprises at least a two dimensional representation of an object. 30. The method of claim 29 wherein the conjecture comprises an image of a previously reconstructed object. 30. The method of claim 29 wherein the conjecture comprises an image of a previously photographed object that is similarly made. 34. The method of any one of the preceding claims, wherein the indication comprises a possible deviation in the desired internal structure. 35. The method of any one of claims 1 to 34, wherein reconstructing the image comprises reconstructing only a portion of the object, in accordance with the instructions. As an image reconstruction method, Providing projection data of a non-animal object; Reconstructing an image from the projection data, Different reconstruction processes are applied to at least a portion of the image, and wherein the different reconstruction processes include different reconstruction methods for the at least a portion. As an image reconstruction method, Providing projection data of a non-animal object; Reconstructing an image from the projection data, Different reconstruction processes are applied to at least a portion of the image, wherein the different reconstruction processes include using different values for at least one reconstruction parameter of the reconstruction method used for the at least a portion. As an image reconstruction method, Providing projection data of a non-animal object; Reconstructing an image from the projection data, Different reconstruction processes are applied to at least a portion of the image, wherein the different reconstruction processes include a method of reconstructing the at least a portion at different spatial resolutions. 39. The method according to any one of claims 36 to 38, wherein said at least part comprises at least two parts, each of which requires different treatments for each other and for the third part of the image. As an image reconstruction method, Providing projection data of a non-animal object; Reconstructing an image from the projection data, Wherein a different reconstruction process is applied to at least a portion and at least a second portion of the image and is applied to the image at least three different processes, one for each portion and at least a different portion of the image. 41. The method of any of claims 36-37 or 40, wherein different reconstruction processes include reconstructing at different resolutions. The method according to any one of claims 36 to 41, Providing an internal structure indication of the object, And wherein said different reconstruction processes are applied according to said instructions. As an image reconstruction method, Providing projection data of a non-animal object; Providing an indication of the internal structure of the object, and Reconstructing an image from the projection data, A different reconstruction method is applied to at least a portion of an image, wherein the different reconstruction method is provided according to the instructions. 44. The method of claim 42 or 43, wherein the step of providing an indication comprises the step of reconstructing using a low quality image of the object. 45. The method of claim 44, wherein the low quality image is reconstructed using a reverse projection method. 45. The method of claim 44 wherein the low quality image is reconstructed using a Bayesian method. 45. The method of claim 44, wherein the low quality image is reconstructed using a Fourier transform method. 45. The method of claim 44 wherein the low quality image is reconstructed using a maximum likeness method. 45. The method of claim 44 wherein the low quality image is reconstructed using a maximum entropy method. 50. The method according to any one of claims 44 to 49, wherein the low quality image is reconstructed using a different resolution chart than that used in the image reconstruction step. 51. The method of any one of claims 42-50, wherein the image reconstruction comprises reconstructing using an algebraic reconstruction method. 51. The method of any one of claims 42-50, wherein the image reconstruction comprises reconstructing using a Bayesian method. 51. The method of any one of claims 42-50, wherein the image reconstruction comprises reconstructing using a Fourier transform. 51. The method of any one of claims 42-50, wherein the image reconstruction comprises reconstructing using a maximum likeness method. 51. The method of any one of claims 42-50, wherein the image reconstruction comprises reconstructing using a maximum entropy method. 51. The method of any one of claims 42-50, wherein the image reconstruction comprises reconstructing using a rear projection method. 51. The method of any one of claims 42-50, wherein image reconstructing comprises reconstructing using a finer resolution than that used for the indication. 58. The method of any one of claims 42-57, wherein said spatial processing comprises setting an initial guess of an image in accordance with said indication. 59. The method of any one of claims 42-58, wherein the step of providing instructions includes recovering instructions generated upon imaging of a similar object. 60. The method of any one of claims 42-59, wherein providing an indication comprises providing a design specification of an object. 61. The method of any one of claims 42-60, wherein providing the instructions comprises providing a fabrication specification of the object. 62. The method of any one of claims 42-61, wherein the spatial processing is based on speculative distances from at least one edge of the at least a portion of the image, wherein the predictive speculation is based on the indication. 63. The method of any one of claims 42-62, wherein the spatial processing is in accordance with the at least part of an expected reconstruction error, wherein the expected error is based on the indication. 64. A method according to any one of claims 42 to 63, wherein said spatial processing method depends on an expected manufacturing error in said at least part, and said expected manufacturing error is based on said indication. 65. The method according to any one of claims 42 to 64, wherein said spatial treatment method depends on the reliability of the internal structure of the object, said reliability being based on said indication. 66. The method of any one of claims 36-65, wherein said at least some is at least one band surrounding the object. Wherein the band includes a region overlapping an outer edge of at least one object. 67. The method of claim 66 wherein the band comprises an area surrounding the opening of the object. 66. The method of any one of claims 36-65, wherein said at least part substantially surrounds a unique feature of the object. 70. The method of claim 69, wherein said feature comprises a region having a particular density. 71. The method of any one of claims 36-70, wherein the spatial processing comprises using speculation about the selected pixel within the at least a portion, instead of reconstructing the selected pixel. 72. The method of any one of claims 36-71, wherein the spatial processing comprises using speculation about the pixel selected outside of the at least a portion, instead of reconstructing the selected pixel. 72. The method of any one of claims 36-71, wherein said spatial processing comprises providing a low quality reconstruction outside of said at least a portion. 74. The method of claim 73 wherein the low quality reconstruction comprises an error reconstruction. 74. The method of claim 73 wherein the low quality reconstruction comprises low spatial resolution reconstruction. 76. The method of any of claims 36-75, wherein the same pretreatment is applied to the interior and exterior of the at least a portion. 76. The method of any one of claims 36-75, wherein different pretreatments are applied to said at least one minute of inner and outer pixels. 78. The method of claim 76 or 77, wherein the pretreatment comprises filtering. 79. The method of any one of claims 36-78, wherein said spatial treatment is in accordance with the level of detail needed in said at least part. 80. The method of any of claims 36-79, wherein the spatial treatment is in accordance with the measurement to be performed on the at least a portion. 81. The method of any one of claims 36-80, wherein the spatial treatment is in accordance with the material composition of the object. 82. The method of claim 81, wherein the object comprises a composite material and the spatial treatment is in accordance with at least one property of the composite material. 83. The method of claim 82, wherein said at least one property comprises a fiber direction of said material. 83. The method of claim 82, wherein said at least one property comprises a cell size of a material. As a method of reconstructing an image, Providing projection data of a non-animal object; Providing an indication of the internal structure of the object and Reconstructing an image from the projection data, wherein the reconstructing step includes only reconstructing pixels from data of at least one predetermined area of the image, wherein the area has a shape determined according to the instruction Image reconstruction method. 86. The method of claim 85, wherein said at least one predetermined zone comprises at least two discrete zones. 87. The method of claim 85 or 86, wherein the shape comprises a band shape surrounding at least one region of the non-recurring pixel. 88. The method according to any one of claims 85 to 87, wherein the shape is determined from a boundary of the object indicated by the indication. As a method of reconstructing an image, Providing projection data of a non-animal object; Providing an indication of the internal structure of the object and Reconstructing an image from the projection data, in accordance with at least one potential problem area in the indication. 89. The method of claim 89, wherein said at least one potential problem area comprises an edge. 91. The method of claim 89 or 90, wherein said at least one potential problem area comprises a suspected crack area. 91. The method of claim 89 or 90, wherein said at least one potential problem area comprises a suspected void area. 93. The method of any of claims 89-92, wherein the reconstruction method comprises varying the spatial resolution of the reconstruction, depending on the location of the at least one latent problem area. 94. The method according to any of claims 89 to 93, wherein the reconstruction comprises changing the gray level resolution of the reconstruction according to the position of the at least one latent problem area. 95. The method of any one of claims 89 to 94, comprising defining to reconstruct an area differently according to the at least one latent problem area. 96. The method of any of claims 89-95, comprising a method of determining a local certainty level according to the edge. As a method of reconstructing a repeated image, Providing projection data of the non-animal object to be photographed; A first reconstruction step for the target object from the projection data; Discarding at least some of the data in accordance with the first reconstruction; and After the discarding step, repeating the first reconstruction step at least once. As a repetitive image reconstruction method, Providing projection data of the non-animal object to be photographed; Generating a reconstruction contrast; Reconstructing the object from projection data according to the reconstruction contrast; Discarding at least some of the contrast in accordance with the first reconstruction; and After the discarding step, repeating the first reconstruction step at least once. As a repetitive image reconstruction method, Providing projection data of the non-animal object to be photographed; Generating a reconstruction contrast; First reconstructing the object from the projection data according to the reconstruction contrast; Generating a relaxation coefficient in accordance with the first reconstruction; and And repeating the first reconstruction using the relaxation coefficient. As a repetitive image reconstruction method, Providing projection data of a non-animal object to be photographed; Generating a reconstruction contrast; First reconstructing the object from the projection data according to the reconstruction contrast; Changing the value of the first reconstruction according to the predominant distribution function; and Repeating the first reconstruction at least once after the modifying step. 101. The method of claim 100, comprising determining the dominant distribution to be zero outside a convex object defined by all traces of projection data having a zero projection value. 102. The method of claim 101 wherein the dominant distribution function comprises at least three values. 107. The method of any one of claims 97 or 102, comprising processing data prior to repeating the first reconstruction. 103. The method of any one of claims 97 or 103, wherein said first reconstruction comprises a plurality of repetitions. 105. The method of claim 97 or 104, wherein the repeated reconstruction includes multiple repetitions. 107. The method of any one of claims 97 or 105, wherein said repeated reconstruction comprises employing different preprocessing for said first reconstruction of said object, in accordance with said first reconstruction. 107. The method of any one of claims 97 or 106, wherein the repeated reconstruction comprises employing different preprocessing for the projection data, in accordance with the first reconstruction. 108. The method according to any one of claims 97 to 107, wherein said repeated reconstruction comprises employing different preprocessing for said projection data, in accordance with knowledge of the internal structure of said object. 109. The method of any one of claims 97-108, wherein the repeated reconstruction comprises employing different preprocessing for the projection data, in accordance with knowledge of the internal structure of the object. As an image acquisition method of a target, Obtaining a projection data set; Reconstructing a first image from the projection data using a first reconstruction method; Analyzing the image to determine special processing for a portion of the image; and Using the analysis, reconstructing a second image of the object, wherein the second reconstruction is a different reconstruction method than the first reconstruction. 117. The method of claim 110, comprising acquiring data for the second configuration in accordance with the analysis. 117. The method of claim 111, wherein the data is acquired according to desired image quality in the second image. 117. The method of claim 111 or 112, wherein said data is acquired in accordance with a desired analysis of said second image. 117. The method according to any one of claims 111 to 113, comprising changing the contrast of the ion radiation used for the data acquisition according to the analysis. 116. The method of any of claims 111-113, comprising changing the wavelength of ion radiation used for the data acquisition in accordance with the analysis. 116. The method of claim 114 or 115, wherein the ion radiation comprises x-rays. 116. The method of claim 114 or 115, wherein the ion radiation comprises gamma rays. 116. The method of any one of claims 111-113, wherein the data is acquired using non-ionized electromagnetic radiation. 118. The method of any of claims 111-118, comprising changing at least one parameter of the detection circuit in accordance with the analysis. 119. The method of claim 119, wherein said at least one parameter is a gain. 121. The method of any one of claims 111 or 120, comprising selecting at least one of the detection systems from a plurality of available devices in accordance with the analysis. 121. The method of claim 121, wherein said device comprises a detector. 123. The method of claim 121, wherein said device comprises a filter. 123. The method of claim 121, wherein said device comprises a collimator. 124. The method of claim 110 or 124, wherein the second reconstruction method comprises an algebraic reconstruction method. 126. The method of claim 125, wherein the algebraic reconstruction method comprises an ART-like method. 124. The method of claim 110 or 124, wherein the second reconstruction method comprises a Bayesian reconstruction method. 124. The method of claim 110 or 124, wherein the second reconstruction method comprises a Fourier transform reconstruction method. 124. The method of any one of claims 110 or 124, wherein the second reconstruction method comprises a maximum co-current reconstruction method. 124. The method of any one of claims 110 or 124, wherein the second reconstruction method comprises a maximum entropy reconstruction method. 124. The method of claim 110 or 124, wherein the second reconstruction method comprises a rear projection reconstruction method. 137. The method of claim 110 or 131, wherein the second reconstruction method uses a different resolution grid than the first reconstruction method. 133. The method of claim 110 or 132, wherein the first method comprises a reverse projection method. 133. The method of any one of claims 110 or 132, wherein the first reconstruction method comprises a Bayesian method. 133. The method of any one of claims 110 or 132, wherein the first reconstruction method comprises a maximum likeness method. 133. The method of any one of claims 110 or 132, wherein the first reconstruction method comprises a maximum entropy method. 133. The method of any one of claims 110 or 132, wherein the first reconstruction method comprises a Fourier transform method. 138. The method of claim 110 or 137, wherein the first reconstruction fails to achieve sufficient convergence for at least a portion of the image. 138. The method of claim 138, wherein the failure is determined from an error level. 138. The method of claim 138, wherein the failure is determined from an unexpected value for the reconstructed pixel value. Obtaining a projection data set; Reconstructing a first image from the projection data using a first reconstruction; Analyzing the image to determine special processing for the image portion; The data acquisition arrangement comprises a device selected from an available set of functionally identical devices, wherein the projection data acquisition stage acquires data for a second reconstruction using a data acquisition arrangement different from the first arrangement used; And reconstructing a second image of the object. 145. The method of claim 141, wherein said different configuration uses a different optical detector than said first configuration. 145. The method of claim 141, wherein the different configuration uses a different filter than the first configuration. 145. The method of claim 141, wherein said different configuration uses a different detector circuit than said first configuration. 145. The method of any one of the preceding claims, comprising post-processing the reconstructed image using an image processing method. 145. The method of claim 145, wherein the image processing method is adapted to enhance feature measurement in the reconstructed image. As a CT imaging device, x-ray radiation source; A data acquisition system comprising at least one set of functionally identical devices, comprising: a data acquisition system for acquiring attenuation data in accordance with the x-ray density of an object disposed in the system; And a controller for selectively selecting a particular one of the devices to be used in the data acquisition system in accordance with its own analysis of data acquired through the system including a different one of the devices. Algebraic CT image reconstruction method, Providing projection data; Generating contrast on the moment of data; and And reconstructing an image from the data using the moment contrast. 148. The method of claim 148, wherein the moment comprises a primary moment. 148. The method of claim 148, wherein said moment comprises a secondary moment. 148. The method of claim 148, wherein the moment comprises a higher order moment than the second moment. Relaxation coefficient generation for iterative algebraic reconstruction Providing a reconstructed previous image by setting a card value during a certain repetition; and And setting a relaxation coefficient for continuous repetition according to a change value between the contrast value and the value in the image. 152. The method of claim 152, wherein the relaxation factor setting step Comparing the contrast value to a threshold value; and Setting a relaxation coefficient according to the comparison. 153. The method of claim 153, wherein said threshold is a statistical principle function of said variation. As a CT image reconstruction method, Providing projection data of the non-animal object from the plurality of projection angles; and And back-projecting the data into a data structure representing a variable resolution grid. And the data structure comprises a hierarchical data structure. As an industrial inspection system, Feeders of one or more non-animal objects to be photographed, and An industrial inspection system mechanically connected to said feeder, said CT imager photographing said at least one object. 169. The industrial inspection system of claim 157 wherein the source comprises a conveyor belt for carrying a plurality of objects. 158. The industrial inspection system of claim 158 wherein the object is assembled from subcomponents. 158. The industrial inspection system of claim 158, wherein the object is molded. 158. The industrial inspection system of claim 158, wherein the object is rolled. 158. The industrial inspection system of claim 158, wherein the object is injection molded. 158. The industrial inspection system of claim 158, wherein the feeder comprises a take-up device. 158. The industrial inspection system of claim 158, wherein the object is mechanized. 158. The industrial inspection system of claim 157, wherein the feeder comprises an extrusion machine for extruding the object in the form of a continuous profile. 167. The industrial inspection system of claim 165, wherein the extruder comprises an extrusion nozzle. 167. The industrial inspection system of claim 165, wherein the extruder comprises a shaper. 167. The industrial inspection system of any of claims 157-167, wherein the object consists essentially of one material. 167. The industrial inspection system of any of claims 157-167, wherein the object consists essentially of two or more materials. 167. The industrial inspection system of any of claims 157-167, wherein the object is comprised of a composite material. 167. The industrial inspection system of any of claims 157-167, wherein the object consists of two or more materials. 172. The industrial inspection system according to any one of claims 157 to 171, wherein the CT imager is tuned to photograph an object according to the supply of the object of the feeder. 172. The industrial inspection system according to any one of claims 157 to 172, wherein the CT camera captures the object using a spatial imaging method. 174. The industrial inspection system according to any one of claims 157 to 173, wherein said CT imager employs a method according to any of claims 1-147. As a CT camera for industrial photography, A detector; x-ray source; Shooting area and CT camera including a manipulator to lift and carry the object to be photographed in the photographing. 175. The CT imager of claim 175, wherein the manipulator comprises a robotic arm. 175. The CT camera of claim 175, wherein said manipulator comprises a winch. As a quality assurance manufacturing method, Manufacturing an object; Photographing the object with a CT camera; Measuring features of the photographed image to detect deviations in the design specification and the internal structure of the object; and Discarding the object if it fails to meet a design specification. 178. The method of claim 178, wherein the object is molded. 178. The method of claim 178, wherein the object is extrusion molded. 178. The method of claim 178, wherein the object is molded. 181. The method of claim 178 or 181, further comprising: analyzing the image to detect at least one material defect and And discarding the object in accordance with the detected defect. 182. The method of claim 182, wherein the at least one defect comprises a bubble. 182. The method of claim 182, wherein the at least one defect comprises a void. 182. The method of claim 182, wherein the at least one defect comprises a change in density. 182. The method of claim 182, wherein the at least one defect comprises a crack. 182. The method of claim 182, wherein the at least one defect comprises a crust. 148. The method of claim 1 or 146, wherein the object comprises a fabricated object. 187. The method of claim 188, wherein the object is a template object. 187. The method of claim 188, wherein the object is a compression molded object. As a tomographic reconstruction method. Acquiring photographic data from a plurality of projection angles of the non-animal object; First reconstructing an image from the photographing data using a tomographic reconstruction method, and A tomographic reconstruction method is provided that includes introducing discretization into the image to convert the image value of the reconstructed image into a limited set of allowed image values, the discretization maintaining at least one image characteristic, At least one image characteristic comprises a moment of the image. 192. The method of claim 191, wherein the moment is a primary moment. 192. The method of claim 191, wherein the moment is a second or higher moment. 191. The method of any one of claims 191 to 193, wherein the limited set value comprises only two values. 194. The tomographic reconstruction method of any of claims 191 to 194, comprising the step of second reconstructing the image, in reciprocal fashion after the step of discretizing introduction. 195. The method of any one of claims 191 to 195, wherein the limited set value is dependent on the number of values and a value relative to the expected density value of the image. As a tomographic reconstruction method. Providing a non-animal object having at least one subcomponent in a known geometry; Obtaining photographing data of the object from a plurality of projection angles; and And reconstructing the image of the object using the at least one subcomponent of known geometry. The method of claim 197, wherein the image reconstruction method is Generating a low quality image reconstruction; Identifying subcomponents of the image; and Using said identification, further reconstructing said image as a principle. 199. The method of claim 198, wherein identifying subcomponents includes identifying positions of the subcomponents. 199. The method of claim 198, wherein identifying subcomponents comprises identifying a direction of the subcomponents. 202. The method of any one of claims 198-200, wherein the further reconstruction includes fixing a value for a pixel of the image, wherein the pixel corresponds to a portion of at least one subcomponent, The fixing step is in accordance with the known geometry. 202. The method of any one of the claims 197-201, wherein the object essentially comprises at least one subcomponent. 202. The method of any one of the claims 197-202, wherein the at least one subcomponent comprises two subcomponents in a known geometry.