KR100972425B1 - Multiple illumination path system and method for defect detection - Google Patents

Abstract

A method is provided for detecting defects by a multiple illumination path system, the method comprising performing an initial scan of an object, wherein during the initial scan a plurality of illumination paths simultaneously illuminate the object; In response to detection signals obtained during the initial scan, dividing the object into segments; For each segment, selecting a selected illumination path to be activated during a defect detection scan of the segment, in response to at least one parameter selected from the dominant material of the segment and the surface roughness level of the segment; Illuminating each segment of the object only by the selected illumination path of the object; And in response to the illumination, generating detection signals indicative of the defects.

Lighting Path, Surface Roughness, Dominant, Defect Detection, Optical Inspection, Scanning

Description

MULTIPLE ILLUMINATION PATH SYSTEM AND METHOD FOR DEFECT DETECTION

Related application

This application claims priority from US Provisional Patent Application No. 60 / 941,672, filed June 3, 2007.

Technical Field

The present invention relates to an automatic optical inspection system.

At the end of the production line, prior to packaging and shipping, the finished High Density Interconnect (HDI) substrate should be visually inspected for surface defects and / or integrity violations. These visual inspection results will classify these substrates as validated or rejected according to Quality Control criteria.

Surface defects are the result of unsatisfactory production quality and / or unsatisfactory handling. Here, the overall quality relates to all included surface types, ie metal surfaces such as solder masks and gold plated interconnect pads on laminate base material. Each type of this surface is affected by certain types of defects as a product of the manufacturing process and the origin material.

From an optical point of view, each material introduces different optical properties that need to be processed as it is being inspected (either manually or via an automatic machine).

There is an increasing need to provide effective defect detection methods and systems.

It is an object of the present invention to provide an effective defect detection method and system which enables stable surface defect detection while minimizing false alarms.

In order to solve the above problem, a method for detecting defects by a multiple illumination path system is provided. The defect detection method includes performing an initial scan of an object, wherein during the initial scan a plurality of illumination paths illuminate the object simultaneously; In response to detection signals obtained during the initial scan, dividing the object into segments; For each segment, selecting a selected illumination path to be activated during a defect detection scan of the segment, in response to at least one parameter selected from the dominant material of the segment and the surface roughness level of the segment; Illuminating each segment of the object only by the selected illumination path of the object; And in response to the illumination, generating detection signals indicative of the defects.

Conveniently, the defect detection method further comprises processing the detection signals to detect the defects.

Conveniently, the defect detection method further comprises receiving segment information, wherein the segmenting step also responds to object information representing object segments.

Conveniently, selecting the selected illumination path to be activated during the defect detection scan of the segment is to illuminate the segment by multiple illumination paths, one illumination path at a time, after determining the dominant material of the segment. The illumination step, wherein different illumination paths are suitable for different possible roughness levels of the segment; Generating detection signals resulting from the illumination; And processing the detection signals to select one illumination path for the segment.

Conveniently, the defect detection method utilizes an illumination path that illuminates the metal surface and the metal surface defects such that the image of the metal surface is substantially white and the image of the metal surface defect is substantially black regardless of the roughness level of the metal surface. The method further includes the step of selecting.

Conveniently, the defect detection method employs an illumination path that illuminates the translucent surface and the translucent surface defects such that the image of the translucent surface is substantially black and the image of the translucent surface defect is substantially white regardless of the roughness level of the translucent surface. The method further includes the step of selecting.

Conveniently, the defect detection method further comprises selecting from a specular illumination path, a diffuse oblique channel, and a near-specular illumination path.

Conveniently, the defect detection method further comprises generating detection signals by an imaging path comprising a camera and a mask, the camera comprising a sensing surface parallel to the surface of the object, wherein the mask is at an angle of the imaging path. Introduce asymmetry in the angular collection range.

Conveniently, the defect detection method further comprises selecting from a specular illumination path, a front oblique illumination path, and a back-scattering oblique illumination path.

Conveniently, the defect detection method further comprises selecting one of the two configurations of the configurable illumination path, wherein in the first configuration, the configurable illumination path illuminates the object in a continuous angle range, In a second configuration, the configurable illumination path illuminates the object in two sub-ranges of continuous angular range spaced from each other.

Conveniently, the defect detection method further comprises selecting from a front oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path, wherein the camera generating the detection signals comprises: an object; The object is imaged at an angle that is substantially different from the normal to its surface.

Conveniently, the defect detection method is adapted to the segment in response to classification of the segment as comprising a shiny metal surface, a semi-matt metal surface, a matt metal surface, or a translucent dielectric surface. Selecting the selected illumination path for the step.

Conveniently, the defect detection method comprises the steps of accommodating a group of objects comprising objects and other objects, the objects and other objects ideally coinciding with each other and undergoing substantially the same manufacturing conditions; And illuminating each segment of each other object by the selected illumination path associated with the corresponding segment of the object.

Conveniently, the object is an electrical circuit, and the defect detection method comprises the steps of: receiving a group of electrical circuits including the object and other objects; And illuminating each segment of each other object by the selected illumination path associated with the corresponding segment of the object.

A defect detection method is provided. This defect detection method illuminates each segment of an object only by a selected illumination path associated with the segment, wherein the single selected illumination path responds to at least one parameter selected from the dominant material of the segment and the roughness level of the segment. An illumination step; In response to illumination, generating detection signals indicative of defects; And processing the detection signals to detect the defects.

Conveniently, the defect detection method further comprises illuminating the segment of the object by a selected illumination path selected from a specular illumination path, a diffuse oblique channel, and a near- specular illumination path.

Conveniently, the defect detection method further comprises generating detection signals by an imaging path comprising a camera and a mask, the camera comprising a sensing surface parallel to the surface of the evaluated object, the mask Introduces asymmetry in the angular acceptance range of the imaging path.

Conveniently, the defect detection method further comprises illuminating the segment by a selected illumination path selected from a specular illumination path, a front oblique illumination path, and a backscatter oblique illumination path.

Conveniently, the defect detection method further comprises selecting one of the two configurations of the configurable illumination path, wherein in the first configuration, the configurable illumination path illuminates the object in a continuous angle range, In a second configuration, the configurable illumination path illuminates the object in two sub-ranges of continuous angular range spaced from each other.

Conveniently, the defect detection method comprises the steps of illuminating a segment by a selected illumination path selected from a front oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path; And generating detection signals by imaging the object at an angle substantially different from the normal to the surface of the object.

Conveniently, the defect detection method further comprises illuminating the segment by the selected illumination path, wherein the selection of the selected illumination path comprises a polished metal surface, a semi-gloss metal surface, a matte metal surface or a translucent dielectric surface. Respond to classifying the segment as one.

Conveniently, the defect detection method further comprises selecting a selected illumination path in response to the examination of the corresponding segment of the other object.

Conveniently, the object and the other object belong to the same batch of electrical circuits.

The invention will be more fully understood and appreciated from the following detailed description taken in conjunction with the accompanying drawings.

According to the present invention, it is possible to detect a stable surface defect while minimizing false alarms.

Multiple illumination path systems are provided. The multiple illumination path system includes a plurality of illumination paths, a processor, and an imaging path. The processor controls (i) the initial scan of the object; (Ii) in response to the detection signals obtained during the initial scan, divide the object into segments; (Iii) determine the surface roughness of each segment; (Iii) for each segment, in response to at least one parameter selected from the dominant material of the segment and the surface roughness level of the segment, it is suitable for selecting the selected illumination path to be activated during the defect detection process of the segment. Multiple illumination paths are suitable for illuminating an object simultaneously during the initial scan. The selected illumination path illuminates a segment associated with the selected illumination path during a defect detection scan of the object. The imaging path is suitable for generating detection signals indicative of defects in response to illumination.

Conveniently, the processor is suitable for processing the detection signals to detect defects.

Conveniently, the processor is suitable for segmenting the object under evaluation in response to object information indicative of the object segments.

Conveniently, the multiple illumination path system (i) illuminates a segment by multiple illumination paths, one illumination path at a time; (Ii) generating, by the imaging path, detection signals resulting from the illumination; (Iii) the processor is adapted to process the detection signals to select one illumination path for the segment, where different illumination paths are suitable for different surface roughness levels.

Conveniently, the processor illuminates the conductor surface and conductor surface defects such that the image of the conductor surface is substantially white and the image of the conductor surface defect is substantially black, regardless of the roughness level of the conductor surface (such as metal conductor). It is suitable for selecting the lighting path.

Conveniently, the processor illuminates the translucent dielectric surface and the translucent dielectric surface defects such that the image of the translucent surface is substantially black and the image of the translucent (dielectric) surface defect is substantially white regardless of the roughness level of the translucent surface. It is suitable for selecting the lighting path.

Conveniently, the plurality of illumination paths include a specular illumination path, a diffuse oblique channel, and a near- specular illumination path.

Conveniently, the imaging path comprises a camera, an objective lens and a mask, the camera comprises a sensing surface parallel to the surface of the object under evaluation, the objective lens imaging the object with the sensing area of the camera, and the mask is an imaging path. Introduce asymmetry in the angle acceptance range.

Conveniently, the plurality of illumination paths include a specular illumination path, a front oblique illumination path, and a backscatter oblique illumination path.

Conveniently, the multiple illumination path system further comprises a configurable illumination path, which can have one of two configurations, wherein in the first configuration, the configurable illumination path illuminates an object in a continuous angular range and In a second configuration, the configurable illumination path illuminates the object in two sub-ranges of continuous angular range spaced from each other.

Conveniently, the plurality of illumination paths include a front oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path, wherein the imaging path points the object at an angle that is substantially different from the normal to the surface of the object. And a camera for imaging.

Conveniently, the processor is suitable for selecting a selected illumination path for the segment in response to the classification of the segment as comprising a polished metal surface, a semi-gloss metal surface, a matte metal surface, or a translucent dielectric surface.

Conveniently, the illumination path is suitable for illuminating a segment of another object by a selected illumination path associated with the corresponding path of the object, the other object ideally matching the object, and the object and the other object being substantially identical Go through the conditions.

Conveniently, the illumination path is suitable for illuminating a segment of another object by a selected illumination path associated with the corresponding segment of the object, the object and the other objects being electrical circuits belonging to the same group.

A defect detection system is provided. This defect detection system includes a processor, an imaging path, and a plurality of illumination paths. The defect detection system is suitable for illuminating each segment of an object only by a selected illumination path associated with the segment, wherein the selected illumination path is in response to at least one parameter selected from the dominant material of the segment and the surface roughness level of the segment. Is selected. The imaging path is suitable for generating detection signals indicative of defects in response to illumination. The processor processes the detection signals to detect the defects.

Conveniently, the plurality of illumination paths include a specular illumination path, a diffuse oblique channel, and a near- specular illumination path.

Conveniently, the imaging path comprises a camera and a mask, the camera comprises a sensing surface parallel to the surface of the object under evaluation, and the mask introduces asymmetry into the angle acceptance range of the imaging path.

Conveniently, the plurality of illumination paths include a specular illumination path, a front oblique illumination path, and a backscatter oblique illumination path.

Conveniently, the defect detection system further comprises a configurable illumination path having two configurations, in the first configuration, the configurable illumination path illuminates the object in a continuous angle range and, in the second configuration, configurable The illumination path illuminates the object in two sub-ranges of continuous angular range spaced from each other.

Conveniently, the plurality of illumination paths include a front oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path, wherein the imaging path points the object at an angle that is substantially different from the normal to the surface of the object. And a camera for imaging.

Conveniently, the processor selects the selected illumination path in response to the classification of the segment as comprising a polished metal surface, a semi-gloss metal surface, a matte metal surface, or a translucent dielectric surface.

Application Immutable Image Appearance and Integrated Defect Detection Algorithm

In order to be consistent and standardized in the defect detection process application, and therefore fast and reliable accordingly, the target gray scale image should be of the following polarity: (i) Regular metal surfaces are defined by high intensity pixels. Represented as “white” expressed; (Ii) the translucent dielectric surface appears as "black" represented by low brightness pixels; (Iii) metal surface defects appear as “black” on a “white” background—local texture deformations are represented by low brightness pixels surrounded by high brightness regular metal textures; (Iii) Dielectric surface and sub-surface defects appear as "white" on a "black" background. Dielectric surface and sub-surface deformations are represented by high brightness pixels surrounded by low brightness regular dielectric textures.

As shown in FIG. 1, for an object (also referred to as an application) comprising a metal and a dielectric portion (eg-BGA), the image is a high brightness "with a" black "defect 14. A high dynamic range gray scale image with a white " GOLD pixel 13, and a low brightness " black " SOLDER mask pixel 11 with a "white "

Figure 1 also shows these defects-solder mask scratch 17, conductor three-dimensional (3D) defect 18, solder mask void 19, foreign material 16 under the solder mask.

The standardized approach described above indicates a predefined image polarity, where metal three-dimensional (3D) defects appear as "black" on "white", whereas surface and sub-surface dielectric defects are "white" on "black". Appears as. In order to create black pixels on the white periphery on 3D defects, shadowing and occlusion effects are induced by the use of specially designed lighting arrangements. In order to visualize dielectric defects as white pixels in the black periphery, scattering techniques are applied.

Surface finish invariant

Multiple illumination path systems have an optical configuration that provides gray scale images according to a standardized approach regardless of various surface finishes. All surfaces defined as optional objects to be inspected by this multiple illumination path system can be classified into the following groups by their reflectance properties: (i) As shown in FIG. 2A, in particular also the light source 210. And a very smooth, glossy opaque surface 212 (polished metal) that reflects most of the incident light in a specular reflection manner, as shown by illumination pattern 214; (Ii) a slightly rough semi-glossy opaque surface 216 (semi-gloss metal), which reflects incident light in a directional diffusion manner, as shown in FIG. ); (Iii) a rough matte opaque surface that reflects the incident light in a fully diffused (Lambertian and isotropic) manner, as shown in FIG. 2C and also, in particular, as light source 210 and illumination pattern 220. 218 matte metal); (Iii) a smooth dielectric transparent and translucent layer 240 that combines surface specular reflection and volume diffuse reflection, as shown in FIG. 2D and also, in particular, as light source 210 and illumination pattern 230. Insulators).

Gray scale images include (i) the angular position of the imaging and illumination optics (also referred to as imaging path and imaging paths) with respect to the inspection surface; (Ii) surface reflection characteristics; (Iii) the numerical aperture and resolution of the imaging optics; (Iii) the angular coverage of the illumination optics; And (iii) a function of the illumination spectrum.

Conveniently, the multiple illumination path system includes an imaging path that includes a single camera (such as, but not limited to CCD) disposed at a fixed angular position with respect to the object.

In such a system, surface invariant image appearance is achieved by providing multiple illumination paths that differ from each other by their illumination angles and are suitable for various surface finishes.

The use of a single camera dramatically reduces the cost of the system. Multiple light paths and single camera systems are less expensive than multiple camera systems.

Using asymmetric illumination and / or imaging paths

To emphasize the shadowing and closing effects for 3D defect visualization on polished metal surfaces, a principle of broken symmetry is enforced. Carefully designed optical construction with inherent asymmetry emphasizes defect contrast with respect to regular surfaces without introducing undesirable gradient effects. Only specular optical constructions with induced asymmetry provide targeted gray scale images and defects highlighted for the shiny metal surface. This approach is illustrated by the following optical configuration.

Various optical configurations that can provide images of black defects located near white (of FIG. 3 (d)) are shown in FIGS. 3 (a), 3 (b) and 3 (c).

3 (a) shows the tilt imaging and illumination optics in the specular angular position. This oblique imaging and illumination optical system is exemplified by the light source 210, the illumination pattern 302 and the camera 310.

3 (b) shows an orthogonal imaging optics with on-axis specular illumination in which asymmetry is induced by the illumination optics (conveniently by a mask). This orthogonal imaging optical system is illustrated with a light source 210, an asymmetrical illumination pattern 312, a beam splitter 320, and a camera 310.

3 (c) shows an orthogonal imaging optics with on-axis specular illumination. Asymmetry is induced by the asymmetric imaging aperture of the mask 330. This orthogonal imaging optical system is illustrated with a light source 210, an asymmetric imaging pattern 332, a beam splitter 320, and a camera 310.

3 (d) shows a desired image obtained by each of the configurations of FIGS. 3 (a) to 3 (c). The desired image includes a black defect (pixel 350) surrounded by a white background (pixel 340).

Defect highlight lighting on semi-gloss and matt metal surfaces: desirable forbidden angle zones

In a fixed imaging optics position (bevel or orthogonal), the defect polarity (black on white, or white on black) is a function of surface finish (roughness level) and illumination angle. The defect changes its polarity under variable illumination angles.

4A shows a 3D surface defect 402, an irregular gloss texture 404, and a regular semi-gloss texture 406 of the semi-gloss surface 400. The irregular gloss texture 404 is imaged by an imaging path that includes a CCD camera and lens (collectively denoted 310). Various illumination paths exist on both sides of the optical axis of the camera, normal to the surface of the object.

These illumination paths include oblique diffuse illumination paths 410 and 420, near- specular diffuse illumination paths 414 and 424 and unwanted illumination paths (also referred to as prohibited angle zones; 412 and 422). 4B, 4C and 4D show various images obtained when illuminating the semi-gloss surface 400.

3D defects and texture deviations on the semi-gloss surface appear as black pixels 430 and 432 on a white background under near- specular diffused illumination (as shown by paths 414 and 424). In low slope illumination (as shown by paths 410 and 420), the same defects reverse their polarity, ie, white pixels 434 and 435 on the black background 444. There are intermediate angular zones 412 and 422 where the same defects appear indistinguishable from their periphery, such as "gray" on "grey" as shown in FIG. 4C (gray pixel 442).

5A shows a 3D surface defect 502, an irregular gloss texture 504, and a regular matte texture 506 of the matte surface 500. The irregular gloss texture 504 is imaged by an imaging path that includes a CCD camera and a lens (collectively denoted 310). Various illumination paths exist on both sides of the optical axis of the camera, normal to the surface of the object.

These illumination paths include oblique diffuse illumination paths 510 and 520, near- specular diffuse illumination paths 514 and 524 and unwanted illumination paths (also referred to as prohibited angle zones; 512 and 522). 5B, 5C and 5D show various images obtained when illuminating the matte surface 500.

Low-tilt illumination 520 and 510 provide for "right" defect appearance (see FIG. 5D), near- specular illumination 514 and 524 cause confusion in defect color (see FIG. 5B), As shown in FIG. 5C, the intermediate angle illuminations 512 and 524 represent “all grays”.

Near- specular angular zones are preferred for semi-gloss surfaces. Matte surfaces are illuminated at low slopes. The middle angle zone is defined as a prohibition for all metal finishes. In addition, combinations of lights that provide the opposite polarity are also prohibited, because this combination of lights causes non-visualization defects.

Defect highlight lighting for glossy translucent dielectric layers

The target image for this type of application is represented by "white" defect pixels on a "black" background. The optimal solution to the aforementioned requirements is low slope lighting. Low-tilt illumination provides a uniform appearance of "black" of the regular glossy surface without specular reflection. As shown in FIG. 6B, surface and sub-surface defects exhibiting strong scattering centers appear as “white”.

FIG. 6 shows white defects (pixels 642 and 644) over a black background (pixel 646 in FIG. 6B). 6A shows a light pattern 610 resulting from illumination of a solder mask surface that includes sub-surface defects that are highlighted in the imaging process.

Angular coverage and surface defect highlighting

The influence of angular coverage on the appearance of an image is shown in FIGS. 7 (a) to 7 (c). As shown by the graph 740 of FIG. 7B (particularly in comparison with the graph 750 of FIG. 7C), wider angular coverage results in a smoother image. The smoothing effect is shown by an image 731 obtained by light directed at an angle α1, an image 732 obtained from light directed at an angle α2, and an average image 733 obtained from these images. As shown, the optical mean of the signal obtained within the angular range is included. CCD camera 310 obtains the averaged image.

The stronger the surface relief variation, the wider the angular coverage is required to achieve the smoothing effect. In order to emphasize the surface defects, it is necessary to suppress regular texture noise (background smoothing) and to increase surface irregularity local contrast. Angular coverage that satisfies the desired angular region for a particular application preserves the required image polarity and suppresses background noise. This makes the surface detection more stable with a minimum false alarm.

Optimal Optical Configuration for Surface Defect Detection

In order to provide an application constant gray scale image, a multi-angle illumination optical configuration is developed.

An orthogonal configuration in accordance with various embodiments of the present invention is shown in FIGS. 8A and 8B. The imaging optics, including the CCD camera and the objective lens (collectively denoted 310), orthogonally view the inspection surface (the CCD camera is parallel to the object so that the optical axis of the CCD camera is orthogonal to the object).

The dynamic imaging masked aperture 810 inserts an asymmetry into the collection angle 811 of the objective lens. The illumination optics are shown in three independent illumination channels:

An on-axis specular illumination path comprising a light source 130, a mask 820 and a beam splitter 830 that can be used for defect detection on a polished metal surface. This on-axis specular illumination path also changes the angle coverage of the full ([-8 °… + 8 °]) to half ([0 °… + 8 °] or [-8 °… 0 °]). Breaking symmetric masked apertures are also included. The former is shown by the illumination pattern 831 of FIG. 8B.

Approximate-specularly symmetrical illumination paths (850 and 860) for defect detection on semi-gloss metal surfaces. As shown by the illumination patterns 861 and 851 in FIG. 8B, [-32 °... -8 °] and [8 ° ...]. 32 °] there is a corresponding angle coverage.

Low-inclined symmetrical illumination paths 840 and 870 for defect detection on and within matt metal surfaces on and under translucent glossy dielectric layers. As shown by the illumination patterns 841 and 871 in FIG. 8C, [-63 °... -47 °] and [47 ° ...]. 63 °], there is a corresponding angular coverage in the angular zone.

The appropriate illumination channel is selected according to the particular application (particularly, per segment, dominant material and associated surface roughness). Any combination of different illumination channels is prohibited.

The system of the present invention is controlled by a processor 801 to which any of the methods 1000, 1100 or 1200 can be applied.

The tilt configuration is shown in FIG. 8C. The imaging optics, including the CCD camera 310 and the objective lens 890, view the inspection surface at a small tilt angle of -16 °. Further, [5 °... Other viewing angles within the range of 25 ° are acceptable. The illumination optics are shown in three independent illumination channels.

The configurable illumination path may be configured to provide an off-axis wide specular and near- specular illumination path for defect detection on polished and semi-gloss metal surfaces positioned at a specular reflection angle of 16 ° relative to the surface normal. This channel includes an angle steering optical element that switches between two operating positions:

A first position dedicated to the polished surface (also referred to as the 0-position; shown as the continuous angle range 892), or a specular reflection position, wherein the steering element is off and [0 °... 32 °], no part of the full angle coverage is concealed.

A second position (also referred to as the 1-position), or approximate-reflective position, dedicated to the semi-gloss surface, wherein the steering element is on, central, and specularly reflective portion of purely [10 ° -22 °] angular coverage Exclude. The active angle coverage is [0 °... 10 °] and [22 ° ...]. 32 °], which satisfies the approximate-square reflection region. These two sub-angle regions of the angular region 892 are labeled 893 and 894 and are spaced apart by zone 895.

Front oblique illumination for defect detection on matte metal surfaces and translucent glossy dielectric layers. The corresponding angular coverage of the illumination path 896 is [8 °... 32 °]. [-32 °... -8 [deg.]], Another (symmetrical) region is included in the optical path 891 and therefore not shown.

Backscattered oblique illumination for defect detection on, under and under translucent glossy dielectric layers. The corresponding angle coverage is [-28 °... -60 °]. This is shown by light pattern 891.

Certain applications, i.e., off-axis specular illumination in the 0-position for polished metal; (Ii) off-axis specular illumination in one-position relative to the semi-gloss metal; (Iii) forward oblique illumination for matt metal; (Iii) Depending on the front slope and backscatter sloped illumination for the translucent glossy dielectric layer, the appropriate illumination channel is selected.

Conveniently, any combination between (iii), (ii), (iii) or (iii) is prohibited during defect detection.

Example Line Scan Illuminator for Automatic Surface Defect Detection on PBGA, CSP, and Other HDI Applications

9A shows an oblique line scan illumination device, which is designed for automatic surface defect detection for various HDI applications.

The illustrated optical system includes (i) 16 ° inclined imaging optics including a linear CCD array 908 and an objective lens 907; (Ii) three linear light sources 901, 902 and 903 (which may be either linear fiber optics or a line LED array). The light source is perpendicular to the scan direction.

Each linear light source is provided arranged in a parallel manner concentrating on a cylindrical optical system, for example a cylindrical lens doublet. Each concentrated cylindrical doublet has: (iii) a configuration defining angular coverage; And (ii) create an application suitable light line width.

All three focused light lines coincide with the inspection surface.

The linear light source 901 with the cylindrical doublet 904 forms a specular light channel 914 inclined 16 ° from the surface normal. The overall angular coverage provided by the cylindrical doublet 904 is ± 16 °. The rotating linear aperture 912 of FIG. 9B produces a light angle steering effect.

The linear light source 902 with the cylindrical doublet 905 forms a front oblique light channel 925 inclined by 55 ° from the surface normal. The overall angular coverage provided by the cylindrical doublet 905 is ± 8 °.

Linear light source 903 with cylindrical doublet 906 forms backscattering oblique light channel 936 that is inclined by -44 ° from the surface normal. The overall angular coverage provided by the cylindrical doublet 906 is ± 16 °.

The specular light channel 914 (of FIG. 9A) with the rotating angle aperture 912 (of FIG. 9B) in the 0-position represents a shiny gold illumination path.

The specular light channel 914 (of FIG. 9A) with the rotating angle aperture 912 (of FIG. 9B) in the 1-position represents a semi-gloss gold illumination path.

The front oblique light channel 925 represents a matte gold illumination path.

Backscattered oblique light channel 936 and front oblique light channel 925 represent a solder-mask illumination path.

Selection of the selected optical path

In order to provide application constant image formation, an automatic surface identification method associated with the multi-angle device of the present invention is developed. A surface identification algorithm is shown in FIG.

The method 1000 of FIG. 10 begins by step 1010 of scanning an object with full illumination (where all illumination paths are activated at the same time).

Step 1010 is followed by step 1020 of generating a segmentation of the object as defined by the gold segment and the solder mask segment. The division responds to the relative dominance of each material in the segment. The size and shape of the segments may be different.

Multiple segments can be obtained.

For each solder mask segment, step 1020 is followed by step 1025 to select an illumination path for illuminating the solder mask.

For each gold segment, step 1020 is followed by step 1030 which stores the location of the gold segment of the gold pixel forming the segment.

Following step 1030, there are four different illumination paths that can fit four different surface roughness levels (in this embodiment), followed by step 1035 of setting the index i between 1 and 4.

For each index i value, steps 1040, 1045 and 1050 are executed.

Step 1040 includes scanning a segment (frame) with the i-th illumination path.

Step 1040 is followed by step 1045 to determine the gold finish parameter for that segment. This can be done by setting a parameter such as {gold standard deviation (STD); Image global contrast}, where the gold STD refers to the standard deviation of the gold surface signal, and the image global contrast is defined as the ratio between the averaged gold signal and the averaged solder signal.

Step 1040 is followed by step 1050 of analyzing the gold finish parameters calculated by the forced criteria. According to the above definition, the enforced criteria include two simultaneous execution conditions, namely minimum gold standard deviation and maximum global contrast.

After the above-described sequence is completed, the method 1000 proceeds to step 1060 (find the optimal gold finish reference illumination channel indicated by i0) to find the optimal illumination path.

Step 1060 is followed by step 1070 of selecting an illumination path that illuminates this segment and corresponding segments during the defect detection sequence (i0 as the optimal illumination channel for a particular application).

11 illustrates a method 1100 of detecting a defect by a multiple illumination path system in accordance with one embodiment of the present invention.

The method 1100 begins with a calibration sequence comprising steps 1110 through 1140 in which illumination paths are selected per segment.

Once the calibration process is complete, it is noted that the selection of the illumination path per segment can be used during the defect detection scan of one or many ideally matched objects.

For example, a group of electrical objects can be inspected by using a selection of illumination paths (per segment) when performing a calibration process for one electrical circuit and then inspecting the other group of electrical circuits. .

Thus, multiple iterations of various steps of the method 1100 may be performed to detect defects in a group of objects. These multiple iterations are ideally matched objects (objects that are expected to match under ideal conditions), including objects (checked during the calibration process) and other objects whose detection of defects benefits from the results of the calibration process. By accepting groups. Objects and other objects undergo substantially the same manufacturing conditions. Therefore, these surfaces must be characterized by substantially the same roughness level.

This acceptance step is followed by a calibration process. Following the calibration process is followed by a defect detection process that is applied across all (or multiple) other objects.

The method 1100 begins by step 1110 of performing an initial scan of an object. During the initial scan, multiple illumination paths can illuminate an object at the same time.

Step 1110 is followed by step 1120 of dividing the object into segments in response to a detection signal obtained during the initial scan. Each segment is characterized by a dominant substance. For example, the segments may comprise primarily metals, or may comprise primarily translucent dielectrics.

According to various embodiments of the present invention, step 1120 may be followed by step 1140, step 1125, and step 1130 of these combinations.

Step 1125 includes receiving segment information (eg, a design file of the object, a previous scan result, etc.). Following step 1125, in response to the segment information, step 1130 is followed to determine the dominant material of each segment.

Step 1130 includes determining the dominant material of each segment.

Following step 1130, after the dominant material of each segment is known, step 1135 may be followed to determine the surface roughness of each segment. Thus, once the governing material is defined, it is necessary to determine which illumination path optimizes the roughness level of the segment.

For example, step 1135 may comprise (i) illuminating the segment by multiple illumination paths, one illumination path at a time, after determining the dominant material of the segment, where different illumination paths have different possible roughness of the segment. Suitable for the level); (Ii) generating a detection signal resulting from the illumination; And (iii) processing the detection signal to select one illumination path for the segment.

Step 1140 includes, for each segment, selecting a selected illumination path to be activated during the defect detection scan of the segment. This choice is responsive to the dominant material of the segment, but may also be responsive to the surface roughness level of the segment.

Following step 1140, a defect detection sequence follows. This defect detection sequence begins by step 1150 illuminating each segment of the object only by the selected illumination path of the object.

Step 1150 is followed by step 1160, in response to illumination, generating a detection signal indicative of a defect.

Following step 1160, storing the detection signal and additionally or alternatively processing the detection signal may be followed by detecting the defect. Defect detection processes are known in the art. Once a standardized image of the object is received, the defect processing process is faster and more robust.

The method 1100 may be applied for any of the above described systems.

Step 1140 illuminates the metal surface and the metal surface defects such that when the segment dominant material is metal, the image of the metal surface is substantially white and the image of the metal surface defect is substantially black regardless of the roughness level of the metal surface. Selecting an illumination path of the segment.

Step 1140 detects the translucent surface and the translucent surface defects such that when the segment dominant material is a translucent dielectric material, the image of the translucent surface is substantially black and the image of the translucent surface defect is substantially white, regardless of the roughness level of the translucent surface. Selecting an illumination path of the segment to illuminate.

Step 1140 can include selecting from a specular illumination path, a diffuse oblique channel, and a near- specular illumination path. This selection may be associated with some of the systems described above, for example the systems of FIGS. 8A and 8B.

Step 1160 can include generating a detection signal by an imaging path including a camera and a mask. The camera has a sensing surface parallel to the surface of the object. The mask introduces asymmetry into the angular acceptance range of the imaging path. Such detection samples may be associated with some of the systems described above, for example the systems of FIGS. 8A and 8B.

Step 1140 may include selecting from a specular illumination path, a front oblique illumination path, and a backscatter oblique illumination path. Samples of such detection may be associated with some of the systems described above, for example the system of FIG. 8C.

Conveniently, step 1140 may include selecting one configuration from two configurations of the configurable lighting path. In a first configuration, the configurable illumination path illuminates the object in a continuous angle range. In a second configuration, the configurable illumination path illuminates the objects in two sub-ranges of continuous angular ranges spaced from each other. Samples of this configurable illumination path are shown in FIGS. 8C and 9A.

Conveniently, step 1140 includes selecting from a front oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path, wherein the camera generating the detection signal comprises: a normal to the surface of the object; Imaging objects at substantially different angles. Samples of such detection may be associated with some of the systems described above, for example the system of FIG. 8C.

Step 1140 may include selecting a selected illumination path for the segment in response to classifying the segment as including a polished metal surface, a semi-gloss metal surface, a matte metal surface, or a translucent dielectric surface.

Conveniently, the initial scan of step 1110 may differ from the defect detection scan of step 1150 by at least one of the following parameters: speed, signal to noise ratio, and resolution.

Conveniently, the scan that determines the roughness level of the surface of step 1135 may differ from the defect detection scan of step 1150 by at least one of the following parameters: speed, signal to noise ratio, and resolution.

In general, the slower the defect detection scan, the higher the resolution and the better the signal-to-noise ratio than the other scans.

12 illustrates a method 1200 in accordance with an embodiment of the present invention.

The method 1200 includes a defect detection process and uses the information (associated with the selection of the illumination path per segment) during the calibration process.

The calibration process and the defect detection process can be applied to the same object, but this is not necessary. The calibration process can be applied to one object while the defect detection process can be applied to one or more other objects. For example, objects and other objects belong to the same group of electrical circuits.

The method 1200 begins by step 1210 of illuminating each segment of an object only by the selected illumination path associated with the segment. The selected illumination path is selected in response to the dominant material of the segment and may also be responsive to the roughness level of the segment.

Step 1210 is followed by step 1220, in response to illumination, generating a detection signal indicating a defect.

Step 1220 is followed by step 1230 of detecting the defect by processing the detection signal.

Step 1210 may include illuminating a segment of the object by a selected illumination path selected from a specular illumination path, a diffuse oblique channel, and a near-specular illumination path.

Step 1220 may include generating a detection signal by an imaging path including a camera and a mask. The camera comprises a sensing surface parallel to the surface of the object under evaluation. The mask introduces asymmetry into the angular acceptance range of the imaging path.

Step 1210 may include illuminating the segment by a selected illumination path selected from a specular illumination path, a front oblique illumination path, and a backscatter oblique illumination path.

Step 1210 may include selecting one configuration from the two configurations of the configurable illumination path. In a first configuration, the configurable illumination path illuminates the object in a continuous angle range. In a second configuration, the configurable illumination path illuminates the objects in two sub-ranges of continuous angular ranges spaced from each other.

Step 1210 may include illuminating the segment by a selected illumination path selected from a front oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path. Step 1220 may include generating a detection signal by imaging the object at an angle that is substantially different from the normal to the surface of the object.

Step 1210 may include illuminating the segment by a selected illumination path selected in response to the classification of the segment as comprising a polished metal surface, a semi-gloss metal surface, a matte metal surface, or a translucent dielectric surface.

While in this detailed description of the invention numerical values are provided for a number of parameters of the invention, as will be apparent to one of ordinary skill in the art, many different values in accordance with different embodiments of the invention. It is noted that this is applicable to these parameters.

It is also noted that the detailed description and drawings of the present invention detail some embodiments of the invention, and that different embodiments of the invention implement a modified manner of implementing the invention.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrates an example of a defect and a desired gray level signal representing such a defect in accordance with one embodiment of the present invention.

2A-2D illustrate various types of surface finish and reflection modes.

3 (a) to 3 (c) show various lighting and imaging schemes according to various embodiments of the present invention, and FIG. 3 (d) shows a defect and its background according to an embodiment of the present invention. A diagram depicting an image.

4A illustrates an illumination and imaging scheme in accordance with one embodiment of the present invention.

4B-4D show images of defects and backgrounds obtained in response to the different illumination of FIG. 4A.

5A illustrates an illumination and imaging scheme in accordance with one embodiment of the present invention.

5B-5D show images of defects and backgrounds obtained in response to the different illumination of FIG. 5A.

6A illustrates an illumination and imaging scheme in accordance with one embodiment of the present invention.

6B shows an image of a defect and its background obtained in response to the different illumination of FIG. 4A.

7 (a) shows illumination and imaging schemes, and various images obtained from different angles in accordance with one embodiment of the present invention.

7 (b) to 7 (c) show detection signals obtained by applying different illuminations.

8A illustrates a multiple illumination path system in accordance with an embodiment of the present invention.

8B illustrates a multiple illumination path system in accordance with an embodiment of the present invention.

8C illustrates a multiple illumination path system in accordance with an embodiment of the present invention.

9A illustrates a multiple illumination path system in accordance with an embodiment of the present invention.

9B-9D illustrate configurable illumination paths in accordance with one embodiment of the present invention.

10 illustrates a method according to one embodiment of the present invention.

11 illustrates a method according to one embodiment of the present invention.

12 illustrates a method according to one embodiment of the present invention.
13 illustrates a method according to one embodiment of the present invention.

[Description of Drawings]

11: Image gray level profile 310: CCD camera and lens

801: Processor 820: Mask

830: beam splitter 840, 870: low slope symmetrical illumination path

850, 860: Approximate-Specular Illumination Path 890: Objective Lens

Claims (44)

A method of detecting defects by a multiple illumination path system, Performing an initial scan of an object, wherein the plurality of illumination paths simultaneously illuminate the object during the initial scan; In response to the detection signals obtained during the initial scan, dividing the object into segments; For each segment, selecting a selected illumination path to be activated during a defect detection scan of the segment, in response to at least one parameter selected from the dominant material of the segment and the surface roughness level of the segment; Illuminating each segment of the object only by the selected illumination path of the object; And In response to the illumination, generating detection signals indicative of the defects. The method of claim 1, Processing the detection signals to detect defects. The method of claim 1, Receiving segment information, And wherein said dividing step also responds to object information indicative of object segments. The method of claim 1, Selecting the selected illumination path to be activated during the defect detection scan of the segment, After determining the dominant material of the segment, illuminating the segment by multiple illumination paths, one illumination path at a time, wherein the different illumination paths are suitable for different possible roughness levels of the segment. ; Generating detection signals resulting from the illumination; And Processing the detection signals to select one illumination path for the segment. The method of claim 1, Selecting an illumination path illuminating the metal surface and the metal surface defect such that the image of the metal surface is substantially white and the image of the metal surface defect is substantially black regardless of the roughness level of the metal surface. Defect detection method. The method of claim 1, Selecting an illumination path illuminating the translucent surface and the translucent surface defect such that the image of the translucent surface is substantially black and the image of the translucent surface defect is substantially white regardless of the roughness level of the translucent surface. Defect detection method. The method of claim 1, And selecting from specular illumination paths, diffuse oblique channels, and near-specular illumination paths. The method of claim 7, wherein Generating detection signals by an imaging path comprising a camera and a mask, The camera comprises a sensing surface parallel to the surface of the object, And the mask introduces asymmetry into the angular collection range of the imaging path. The method of claim 1, And selecting from the specular illumination path, the front oblique illumination path, and the back-scattering oblique illumination path. The method of claim 1, Further selecting a configuration among the two configurations of the configurable lighting path, In a first configuration, the configurable illumination path illuminates the object in a continuous angle range, In a second configuration, the configurable illumination path illuminates the object in two sub-ranges of the continuous angular range spaced from each other. The method of claim 1, Selecting from a forward oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path, And the camera generating the detection signals images the object at an angle substantially different from the normal to the surface of the object. The method of claim 1, In response to classification of the segment as comprising a shiny metal surface, a semi-matt metal surface, a matt metal surface, or a translucent dielectric surface, selecting a selected illumination path for the segment. Further comprising a defect detection method. The method of claim 1, Receiving a group of objects including the object and other objects, wherein the object and the other objects are ideally matched to each other and have undergone substantially the same manufacturing conditions; And Illuminating each segment of each other object by a selected illumination path associated with the corresponding segment of the object. The method of claim 1, The object is an electrical circuit, The defect detection method, Accepting a group of electrical circuits comprising the object and other objects; And Illuminating each segment of each other object by a selected illumination path associated with the corresponding segment of the object. As a defect detection method, Illuminating each segment of an object only by a selected illumination path associated with the segment, wherein a single selected illumination path comprises at least one parameter selected from the dominant material of the segment and the roughness level of the segment Selected in response to said illuminating step; In response to the illumination, generating detection signals indicative of defects; And Processing the detection signals to detect defects. The method of claim 15, Illuminating the segment of the object by a selected illumination path selected from among specular illumination paths, diffuse oblique channels, and near-specular illumination paths. . The method of claim 15, Generating the detection signals by an imaging path comprising a camera and a mask, The camera comprises a sensing surface parallel to the surface of the evaluated object, And the mask introduces asymmetry into the angular collection range of the imaging path. The method of claim 15, Illuminating the segment by a selected illumination path selected from a specular illumination path, a front oblique illumination path, and a back-scattering oblique illumination path. The method of claim 15, Further selecting a configuration among the two configurations of the configurable lighting path, In a first configuration, the configurable illumination path illuminates the object in a continuous angle range, In a second configuration, the configurable illumination path illuminates the object in two sub-ranges of the continuous angular range spaced from each other. The method of claim 15, Illuminating the segment by a selected illumination path selected from a forward oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path; And Generating the detection signals by imaging the object at an angle substantially different from the normal to the surface of the object. The method of claim 15, Illuminating a segment by the selected illumination path, The selection of the selected illumination path is in response to classifying the segment as comprising a shiny metal surface, a semi-matt metal surface, a matt metal surface or a translucent dielectric surface. Way. The method of claim 15, In response to inspecting a corresponding segment of another object, selecting the selected illumination path. The method of claim 22, The object and the other object belong to the same batch of electrical circuits. As a multiple illumination path system, Multiple lighting paths; A processor; And Includes an imaging path, The processor comprising: Control the initial scan of the object, In response to the detection signals obtained during the initial scan, dividing the object into segments; Determine the surface roughness of each segment, For each segment, suitable for selecting a selected illumination path to be activated during the defect detection process of the segment, in response to at least one parameter selected from the dominant material of the segment and the surface roughness level of the segment, The plurality of illumination paths are suitable for simultaneously illuminating the object during the initial scan, The selected illumination path illuminates a segment associated with the selected illumination path during a defect detection scan of the object, The imaging path is suitable for generating detection signals indicative of defects in response to the illumination. The method of claim 24, And the processor is adapted to process the detection signals to detect defects. The method of claim 24, And the processor is suitable for segmenting an evaluated object in response to object information indicative of object segments. The method of claim 24, The multiple illumination path system, Illuminate the segment by the plurality of illumination paths, one illumination path at a time, The imaging path generates detection signals resulting from the illumination, Suitable for processing, by the processor, the detection signals to select one illumination path for the segment, Different illumination path systems are suitable for different surface roughness levels. The method of claim 24, The processor selects an illumination path that illuminates the metal surface and the metal surface defect so that the image of the metal surface is substantially white and the image of the metal surface defect is substantially black regardless of the roughness level of the metal surface. Suitable, multiple lighting path system. The method of claim 24, The processor selects an illumination path that illuminates the translucent surface and the translucent surface defect so that the image of the translucent surface is substantially black and the image of the translucent surface defect is substantially white regardless of the roughness level of the translucent surface. Suitable, multiple lighting path system. The method of claim 24, The plurality of illumination paths include a specular illumination path, a diffuse oblique channel, and a near-specular illumination path. 31. The method of claim 30, The imaging path comprises a camera and a mask, The camera comprises a sensing surface parallel to the surface of the object under evaluation, And the mask introduces asymmetry into the angular collection range of the imaging path. The method of claim 24, Wherein the plurality of illumination paths comprises a specular illumination path, a front oblique illumination path, and a back-scattering oblique illumination path. The method of claim 24, Further comprising a configurable lighting path that may have one of two configurations, In a first configuration, the configurable illumination path illuminates the object in a continuous angle range, In a second configuration, the configurable illumination path illuminates the object in two sub-ranges of the continuous angular range spaced from each other. The method of claim 24, The plurality of illumination paths include a forward oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path, And the imaging path comprises a camera for imaging the object at an angle substantially different from the normal to the surface of the object. The method of claim 24, The processor is configured to select a segment for the segment in response to classification of the segment as comprising a shiny metal surface, a semi-matt metal surface, a matt metal surface, or a translucent dielectric surface. Multiple light path systems, suitable for selecting light paths. The method of claim 24, The selected illumination path is suitable for illuminating a segment of another object by the selected illumination path associated with the corresponding path of the object, The other object ideally matches the object, and the object and the other object have experienced substantially the same manufacturing conditions. The method of claim 24, The selected illumination path is suitable for illuminating a segment of another object by the selected illumination path associated with the corresponding segment of the object, The object and the other object are electrical circuits belonging to the same group. As a defect detection system, A processor; Imaging paths; And Include multiple lighting paths, The defect detection system is suitable for illuminating each segment of an object only by a selected illumination path associated with the segment, The selected illumination path is selected in response to at least one parameter selected from a dominant material of the segment and a surface roughness level of the segment, The imaging path is suitable for generating detection signals indicative of defects in response to the illumination and And the processor processes the detection signals to detect defects. 39. The method of claim 38, The plurality of illumination paths include a specular illumination path, a diffuse oblique channel, and a near-specular illumination path. 39. The method of claim 38, The imaging path comprises a camera and a mask, The camera comprises a sensing surface parallel to the surface of the evaluated object, And the mask introduces asymmetry into the angular collection range of the imaging path. 39. The method of claim 38, The plurality of illumination paths include a specular illumination path, a front oblique illumination path, and a back-scattering scattering oblique illumination path. 39. The method of claim 38, Further comprising a configurable lighting path having two configurations, In a first configuration, the configurable illumination path illuminates the object in a continuous angle range, In a second configuration, the configurable illumination path illuminates the object in two sub-ranges of the continuous angular range spaced from each other. 39. The method of claim 38, The plurality of illumination paths include a forward oblique illumination path, a backscatter oblique illumination path, and an off-axis specular illumination path, And the imaging path comprises a camera for imaging the object at an angle substantially different from the normal to the surface of the object. 39. The method of claim 38, The processor selects the selected illumination path in response to the classification of the segment as comprising a shiny metal surface, a semi-matt metal surface, a matt metal surface, or a translucent dielectric surface. Fault detection system.

Images