JPH0992195A - Image analysis method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an image analysis method capable of analyzing a image without flatly processing an object face even if a recess or protrusion exists with an observation object and capturing an necessary image for analysis only by moving an object lens in a specified range only once. SOLUTION: Plural pieces of image data V1 to Vn are obtained while changing a distance from a static object to a camera, image data including image data from the protrusion to the recess of the static object is synthesized to make a piece of image data V from among these plural pieces of image data, this one piece of image data is subject to image processing, and image analysis is performed. Change of plural pieces of image data is serially detected, only image data of the changed part is left, the remaining image data is synthesized to be one piece of image data, this one piece of image data is subject to image processing, and image analysis may be performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image analysis method which is convenient when, for example, an image of a stationary object is captured and analyzed by an electron microscope or the like.

[0002]

2. Description of the Related Art As a method of capturing an image of a stationary object with an electron microscope and analyzing the image, a camera mounted on the electron microscope scans the stationary object to obtain image data, and the image data is subjected to comprehensive digital image processing. Device (ID
There is a method of analysis by IP). When capturing an image of a stationary object with an electron microscope, the objective lens of the electron microscope is moved from the state in which it is closest to the stationary object as the sample to the focus position while gradually moving away from it, and the state in which the electron microscope is focused on the stationary object The image is taken in and this image data is used for image analysis.

[0003]

According to the above-mentioned conventional image analysis method, a sharp image cannot be obtained over the entire screen unless the observation target surface of the stationary object is flat.
An object with irregularities cannot be subjected to image analysis, and if an object with irregularities is forcibly observed, it is necessary to process the object surface flat for observation, and the preparation is troublesome. However, it has a drawback that it is not suitable for observing living things. In addition, as described above, the objective position is searched for by moving the objective lens from the state of being closest to the stationary object to the focus position while gradually moving away from it, but it is necessary to repeat the movement of the lens as described above. It is possible to find the focus position for the first time by
The difficulty is in finding the focus position. Further, there is a possibility that the lens may come into contact with the target object and destroy the target object while the lens is repeatedly moved.

The present invention has been made in order to solve the above problems of the prior art, and it is possible to analyze an image of an observation object without processing the observation object surface flat even if the observation object has irregularities. Provide an image analysis method that makes it possible to capture an image required for analysis by moving the objective lens only once within a predetermined range, thereby speeding up image capture and improving operability. The purpose is to do.

[0005]

In order to achieve the above object, the image analysis method according to claim 1 obtains a plurality of image data while changing the distance from the stationary object to the camera.
Image data including image data from a convex to a concave of a static object is combined from the plurality of pieces of image data to form one piece of image data, and this one piece of image data is subjected to image processing for image analysis. It is characterized by performing. The image data of multiple sheets includes image data in which the convex surface of a stationary object is in focus and image data in which the concave surface of a stationary object is in focus.Image data in which the convex surface is in focus and concave surface are in focus The image data including the matched image data is combined into one image data. A static object can be analyzed by processing this one piece of image data.

According to a second aspect of the image analysis method, a plurality of pieces of image data are obtained while changing the distance from the stationary object to the camera, changes in the plurality of pieces of image data are sequentially detected, and only the image data of the changed portion is acquired. And the remaining image data is combined to form one image data, and the image data is subjected to image processing for image analysis. The image data of the changed portion is the image data from the concave to the convex of the object, and the image data from the concave to the convex is left.
It is combined as one piece of image data and used for analysis of stationary objects.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Various embodiments of an image analysis method according to the present invention will be described below with reference to the drawings. The present invention can be implemented using an integrated digital image processing device (IDIP) capable of performing wide-range and multi-purpose image processing. Therefore, first, an outline of the integrated digital image processing apparatus will be described. The material of the image input to the integrated digital image processing apparatus is mainly a camera or video tape recorder (VTR) input image, and not only a digital image but also an analog image can be input. In addition, CD-ROM (read-only memory using compact disc), HDD (hard disk), MO (magneto-optical disc), DVD (digital
It is also possible to have a recording or editing function by using a storage medium such as a video disk), and it is also possible to perform processing such as moving images.

As a structure of the integrated digital image processing apparatus, a workstation or a personal computer is used for performing main processing or sub processing, and a hardware sub processing unit is placed in the main body. The main body includes an input / output unit for material images, a semiconductor memory unit of several tens of Mbytes, an image processing unit including a DSP (digital signal processor) and ASIC (IC for a specific application), and a control unit.

FIG. 8 schematically shows an example of the configuration of the integrated digital image processing apparatus, in which the workstation 4 is placed as the main processing unit and the main body 1 is placed as the sub processing unit.
A material image is input from the CCD camera 2 or the VTR 3 and is capable of performing wide range image processing. The main body 1 may be provided with an interface such as HDD, MO, DVD.

Image analysis using the above-described integrated digital image processing apparatus is performed by the procedure shown in the left half of FIG. In FIG. 7, first, an image is captured in frame units and stored in the image memory. The image samples stored in the image memory are appropriately added and combined according to a capture instruction. When the image capturing is completed in this way, a plurality of combined image data are then subjected to image analysis and image conversion, followed by image correction and final image correction, and output as one image.

The right half of FIG. 7 shows the processing method applied to each stage of the image analysis. First, the captured image data or the image data stored in the image memory is
Image analysis is performed by Laplacian and frequency conversion, and other methods to obtain analysis data. Based on this analysis data, the above-mentioned acquisition instruction is issued, and image samples are added and combined. In image analysis and image conversion after image composition, the data obtained by the analysis data is used as auxiliary data,
Image analysis is performed by edge detection using a method such as Laplacian. In addition, methods such as Hough transform and Fourier transform are used as necessary. The image correction is to correct discontinuous data generated at the boundary between focused images into continuous data. The final image correction is performed by histogram correction by statistical analysis. In the edge detection, Hough transform, Fourier transform, and histogram correction by statistical analysis, coefficients are set and weighting is appropriately performed.

The image analysis method according to the present invention can be carried out by using the integrated digital image processing apparatus as described above. FIG. 1 conceptually shows an image analysis method according to the present invention. For example, while obtaining image data while scanning a stationary object with a camera mounted on an electron microscope, while changing the distance from the stationary object to the camera. Specifically, in the case of an electron microscope, images are taken at regular intervals while sequentially moving away the objective lens from the closest position, and a plurality of image data V1 to Vn are obtained. The image data of V1 to Vn are appropriately combined to form one piece of combined image data V, and the combined image data V is used for image analysis or image processing.

FIG. 3 shows the operation from the image capturing to the image composition. First, the lens is driven so as to cover the in-focus range from the concave portion to the convex portion of the object,
During that time, images are captured (sampling) at regular time intervals. Each sampled image is taken into the IDIP device and stored in the memory, and image data including image data from the concave portion to the convex portion of the object is combined from the stored images to form one image data. Incidentally, image data of, for example, 30 frames can be captured per second. The single combined image data may be a combination of all sampled image data, or an image of a portion necessary for image analysis, that is, a portion including a concave portion to a convex portion of an object. It may be a combination of data.

In the case of the image processing described above, it is necessary to temporarily store all the image data in the memory in the IDIP device and perform the image processing on the stored image data. Therefore, it is not possible to predict how much capacity is required from the image of the concave portion of the object to the image of the convex portion, and the memory capacity may be insufficient.

Therefore, assuming that the memory has a capacity of 60 frames, for example, 30 frames of image data can be captured per second, and it takes 3 seconds to move the camera relative to the object within a predetermined range. If this happens, the memory will be insufficient for 30 frames, so a method of thinning out the frames at regular intervals to 60 frames can be considered. If this method is adopted,
Only the data on the calculation shown by the broken line B in the circle shown in FIG. 2 can be taken out, and the data of the portion necessary for image analysis as shown by the solid line A inside the circle can be taken out faithfully. Can not. Therefore, although it is not possible to obtain a complete processing result, the value between one sampling data and the next sampling data is not so different from continuous data even if an approximate value before and after is taken. , There are no major defects.

However, according to the above method, since it is not processed with accurate data, it is desirable to devise so that it is processed with accurate data by supplementing the above calculated data. Embodiments devised for this purpose are shown in FIGS. 4, 5 and 6. This is one in which changes in image data are sequentially detected, data in a changing portion is sampled, and only necessary image data is fetched. As shown in FIG. 6, first, the lens is driven so as to cover the in-focus range from the concave portion to the convex portion of the object, the images are sampled during that time, and each sampled image is taken into the IDIP device and stored in the memory. Remember. Detect changes in image data from one sampled image data and the next sampled image data,
It is determined whether or not the peak value of the differential density is the highest level based on the change in the image data. If it is the highest level, image data for several frames centering on the peak of the highest level is left and other image data is erased from the image memory,
The remaining images of several frames are combined to form one piece of image data, which is used for image analysis.

Such sampling can be called a real-time sampling method or a random sampling method. According to this method, old data and unnecessary data can be discarded, so that memory can be used. There is an advantage that you can avoid running out of memory without wasting your memory.

FIG. 4 shows an example of image data sampled by the normal sampling and the random sampling method described above, in which (a) is normal sampling and (b) is random sampling. is there.
Both (a) and (b) show changes in the differential density (changes in frequency components) with respect to the sampling time, and the numbers attached along the time axis t indicate the memory numbers stored at the sampling points. And the vertical axis represents the differential concentration. In the normal sampling of FIG. 4A, the dotted line and the solid line are the actual sampling points.

The actual sampling points shown by the dotted line and the solid line in FIG. , ", #, $ Are numbered, a change in data is detected in order from the sampling point! And stored in each memory, an example of memory allocation is shown in FIG. So, the peak value is the highest level at the 19th sampling point.
In the random sampling, as described above, only the image data for several frames centering on the peak of the highest level of the differential density is left. FIG. 4B shows the image data left by such random sampling. Images of several frames left by random sampling are combined into one piece of image data, which is used for image analysis. If the image data sampled by the sampling method is subjected to image analysis, the data of the density change portion can be captured more accurately than in the above-described embodiment, and accurate image analysis using a realistic image can be performed. It can be performed.

According to the embodiment of the present invention described above, a plurality of image data including concave portions and convex portions of an object are combined as one piece of image data. It becomes flat image data without stereoscopic effect or perspective. However, flat image data can be used for image analysis using the IDIP device.

The image analysis method according to the present invention can be applied not only to the analysis of image data captured by an electron microscope as described above, but also to the analysis of image data captured by a general optical microscope. It can also be applied to the field of. For example, even if you try to remotely monitor a monitor panel with many meters arranged through a TV camera, the distance between each meter and the camera is different, so it is not possible to focus the camera on all meters. Impossible, and previously impossible. But,
If the present invention is applied, the camera is moved periodically (the time interval is arbitrarily set) within a range in which all the meters are in focus, and the necessary image data is acquired from among the image data captured during that time. By sampling and synthesizing only the image data, it is possible to obtain image data that is in focus on all the meters, so it is possible to use a monitoring panel with many meters arranged for remote monitoring through a TV camera. It will be possible.

[0022]

According to the first aspect of the present invention, a plurality of pieces of image information are obtained while changing the distance from the stationary object to the camera, and from the plurality of pieces of image data, from the convex to the concave of the stationary object. The image data including the image data is combined into one image data, and this one image data is subjected to image processing for image analysis. It is possible to capture a sharp image from the concave portion to the convex portion as a sharp image without subjecting it to flattening, and to use it for image analysis.
Even organisms can be used for observation. Also, since the image required for image analysis can be captured by moving the camera or objective lens once within a predetermined distance range with respect to a stationary object, it is said that the focus position is searched by repeatedly moving the camera or objective lens. Troublesome operation is unnecessary and the operability is improved. As in the case of the electron microscope, there is less fear that the lens may contact the object and destroy the object while repeatedly moving the lens.

According to the second aspect of the invention, a plurality of pieces of image data are obtained while changing the distance from the stationary object to the camera, and changes in the plurality of pieces of image data are sequentially detected so that only the image data of the changed portion is obtained. And the remaining image data are combined to form one image data, and this one image data is subjected to image processing for image analysis, and therefore, the same effect as the invention according to claim 1 is obtained. In addition, since old data and unnecessary data can be discarded, it is possible to avoid wasting memory and avoid running out of memory. Moreover, since the data of the changed portion can be captured with high accuracy, it is possible to perform accurate image analysis using a realistic image.

[Brief description of drawings]

FIG. 1 is a schematic diagram sensuously showing an image analysis method according to the present invention.

FIG. 2 is a diagram showing an example of image data sampling based on an embodiment of an image analysis method according to the present invention.

FIG. 3 is a flowchart showing an image analysis method according to the embodiment.

FIG. 4 is a diagram showing an example of normal image data sampling and image data sampling based on another embodiment of the image analysis method according to the present invention.

FIG. 5 is a conceptual diagram showing an example of memory allocation by normal data sampling.

FIG. 6 is a flowchart showing an image analysis method according to another embodiment.

FIG. 7 is a flowchart showing an example of an image processing operation of the integrated digital image processing apparatus that can be used for implementing the present invention and a processing method used for the image processing operation.

FIG. 8 is a front view showing an example of the integrated digital image processing apparatus.

Claims (2)

[Claims] 1. A plurality of image data is obtained while changing a distance from a stationary object to a camera, and image data including image data from a convex to a concave of a stationary object is synthesized from the plurality of image data. The image analysis method is characterized in that the image data is converted into one image data, and the image data is subjected to image processing for image analysis. 2. A plurality of pieces of image data are obtained while changing the distance from a stationary object to the camera, changes in the plurality of pieces of image data are sequentially detected, and only the changed image data is left, and the remaining image data is An image analysis method comprising synthesizing one image data and subjecting this one image data to image processing to perform image analysis.