Search

CN-121998968-A - Automatic alignment method and system for light and dark field images of WireBond AOI flyshots, electronic equipment and computer readable medium

CN121998968ACN 121998968 ACN121998968 ACN 121998968ACN-121998968-A

Abstract

The invention discloses a method and a system for automatically aligning bright-dark field images for WireBond AOI flyshots, electronic equipment and a computer readable medium, wherein the method comprises the steps of collecting bright-field reference images and dark-field reference images before flyshots, establishing a shared light field public reference coordinate system, respectively constructing a bright-field characteristic template and a dark-field characteristic template based on the reference images, collecting bright-field images and dark-field images in the flyshots and extracting gradient characteristics to obtain an edge point set, respectively matching the characteristic points of the flyshots with the corresponding templates, calculating direction consistency and space consistency scores, solving respective optimal affine transformation matrixes, combining the optimal affine transformation matrixes based on a pre-calibrated fixed light field offset matrix, calculating a final transformation matrix of a dark-field flyshot image Ji Zhi bright-field coordinate system, respectively applying the final transformation matrix to the bright-field flyshot images and the dark-field flyshot images, and outputting aligned image pairs. The method can respectively carry out the alignment of Ji Mingchang and the dark field, and ensure that the aligned images share the same coordinate system.

Inventors

  • LI GUANGTAO

Assignees

  • 上海铭沣科技股份有限公司

Dates

Publication Date
20260508
Application Date
20260306

Claims (10)

  1. 1. A method for automatically aligning bright-dark field images for WireBond AOI fly-shots, comprising the steps of: Before the start of the fly shooting, acquiring a bright field reference image and a dark field reference image, and establishing a light field common reference coordinate system shared by the bright field reference image and the dark field reference image; Respectively constructing a bright field characteristic template and a dark field characteristic template based on the bright field reference image and the dark field reference image; In the process of flying shooting, acquiring a bright field flying shooting image and a dark field flying shooting image, and respectively extracting gradient characteristics of the bright field flying shooting image and the dark field flying shooting image to obtain a corresponding edge point set; Matching the feature point sets of the bright field aerial image and the dark field aerial image with corresponding templates respectively, calculating the scores of the directional consistency and the spatial consistency, and solving respective optimal affine transformation matrixes; based on a fixed light field offset matrix between reference images and combined with the optimal affine transformation matrix, calculating a final transformation matrix of a light field coordinate system of the pair Ji Zhi of dark field aerial images; And respectively applying the final transformation matrix to the bright field aerial image and the dark field aerial image, and outputting an aligned bright-dark field image pair.
  2. 2. The method for automatic alignment of a light and dark field image for WireBond AOI aerial shots of claim 1 wherein said constructing a bright field feature template and a dark field feature template comprises the steps of: Extracting a training area and a shielding area, and determining an effective template area; calculating gradient amplitude and direction in the effective template area, and extracting edge points; And constructing an image pyramid, and generating a bright field characteristic template and a dark field characteristic template.
  3. 3. The method for automatically aligning bright-dark field images for WireBond AOI flyshots of claim 1, wherein the direction consistency score is calculated based on gradient direction included angles between template points and flyshots, the spatial consistency score is calculated based on spatial errors after affine transformation mapping, and the comprehensive score is a weighted sum of the template points and the flyshots.
  4. 4. The method of claim 1, wherein the fixed light field offset matrix is pre-calibrated based on a spatial transformation relationship between a bright field reference image and a dark field reference image.
  5. 5. An automatic alignment system for a WireBond AOI fly-captured bright-dark field image, comprising: The reference module is used for acquiring a bright field reference image and a dark field reference image before the start of the fly shooting, and establishing a light field common reference coordinate system shared by the bright field reference image and the dark field reference image; the template construction module is used for respectively constructing a bright field characteristic template and a dark field characteristic template based on the bright field reference image and the dark field reference image; The acquisition and extraction module is used for acquiring the bright field aerial image and the dark field aerial image in the aerial shooting process and respectively extracting gradient characteristics of the bright field aerial image and the dark field aerial image to obtain a corresponding edge point set; The computing and solving module is used for respectively matching the feature point sets of the bright field aerial image and the dark field aerial image with the corresponding templates, computing the scores of the direction consistency and the space consistency and solving the respective optimal affine transformation matrixes; the matrix calculation module is used for calculating a final transformation matrix of the bright field coordinate system of the pair Ji Zhi of dark field aerial images based on the fixed light field offset matrix between the reference images and combining the optimal affine transformation matrix; and the image output module is used for respectively applying the final transformation matrix to the bright field aerial image and the dark field aerial image and outputting an aligned bright-dark field image pair.
  6. 6. The automatic alignment system for WireBond AOI aerial photo bright-dark field images of claim 5, wherein said constructing bright-field and dark-field feature templates comprises the steps of: Extracting a training area and a shielding area, and determining an effective template area; calculating gradient amplitude and direction in the effective template area, and extracting edge points; And constructing an image pyramid, and generating a bright field characteristic template and a dark field characteristic template.
  7. 7. The automated bright-dark field image alignment system for WireBond AOI flyshots of claim 5, wherein the direction consistency score is calculated based on gradient direction angles of template points and flyshots, the spatial consistency score is calculated based on spatial errors after affine transformation mapping, and the composite score is a weighted sum of the two.
  8. 8. The automatic alignment system for WireBond AOI-fly-by bright-dark field images of claim 5, wherein the fixed light field offset matrix is pre-calibrated based on a spatial transformation relationship between a bright-field reference image and a dark-field reference image.
  9. 9. An electronic device comprising a memory, a processor, and a computer program stored in the memory, the processor running the computer program to perform a method for automatically aligning a shading field image for WireBond AOI flyshots as set forth in any one of claims 1-4.
  10. 10. A computer readable medium having non-volatile program code executable by a processor, the program code causing the processor to run a shading image auto-alignment method for WireBond AOI flyshots as claimed in any one of claims 1-4.

Description

Automatic alignment method and system for light and dark field images of WireBond AOI flyshots, electronic equipment and computer readable medium Technical Field The present invention relates to the field of machine vision and semiconductor package inspection (AOI) technology, and in particular, to a method and system for automatically aligning bright-dark field images for WireBond AOI flyshots, an electronic device, and a computer readable medium. Background In the Wire Bond process, the AOI equipment needs to detect different types of defects such as gold Wire radian, wire collapse, short circuit, pollution, scratch, edge breakage and the like. In order to improve the production takt, a camera usually adopts a high-speed fly-shooting imaging mode which synchronously moves along with a moving platform, and multi-frame image acquisition is continuously completed under the condition of no stop. However, the following problems exist in the fly-by imaging process: 1. Timing offset of camera exposure and stage motion Because of factors such as movement speed, trigger delay, vibration and the like, unpredictable tiny deviations exist in imaging centers of different frames of images, and the images are directly inconsistent. 2. The bright and dark field light sources cannot be imaged synchronously, and the image characteristics are completely different The bright field image is suitable for gold thread geometric structures; the dark field image is suitable for the defects of surface scratch, edge breakage and the like. The two images have great differences in structural characteristics such as texture, brightness, gradient direction and the like, so that the bright field image is difficult to directly match with the dark field image. 3. Unified compensation model based on mechanical position can not solve pixel offset caused by light source difference Additional displacements are caused by the light source angle, reflection path and exposure time differences, which the mechanical encoder cannot reflect. Traditional image registration methods (such as gray level cross correlation, optical flow, single template matching, etc.) cannot align bright dark field images simultaneously under the condition of fly-swatting. Thus, there is a great need for an accurate method that can separately align Ji Mingchang and dark fields while ensuring that the aligned images share the same coordinate system. Disclosure of Invention According to a first aspect of the embodiment of the present invention, there is provided a method for automatically aligning bright-dark field images for WireBond AOI flyshots, including the steps of: Before the start of the fly shooting, acquiring a bright field reference image and a dark field reference image, and establishing a light field common reference coordinate system shared by the bright field reference image and the dark field reference image; respectively constructing a bright field characteristic template and a dark field characteristic template based on the bright field reference image and the dark field reference image; In the process of flying shooting, acquiring a bright field flying shooting image and a dark field flying shooting image, and respectively extracting gradient characteristics of the bright field flying shooting image and the dark field flying shooting image to obtain a corresponding edge point set; Matching the feature point sets of the bright field aerial image and the dark field aerial image with corresponding templates respectively, calculating the scores of the directional consistency and the spatial consistency, and solving respective optimal affine transformation matrixes; Based on a fixed light field offset matrix between reference images and combined with an optimal affine transformation matrix, calculating a final transformation matrix of a light field coordinate system of the pair Ji Zhi of dark field aerial images; And respectively applying a final transformation matrix to the bright field aerial image and the dark field aerial image, and outputting an aligned bright-dark field image pair. Further, constructing a bright field feature template and a dark field feature template comprises the following steps: Extracting a training area and a shielding area, and determining an effective template area; calculating gradient amplitude and direction in the effective template area, and extracting edge points; And constructing an image pyramid, and generating a bright field characteristic template and a dark field characteristic template. Further, the direction consistency score is calculated based on the gradient direction included angle between the template point and the flying point, the space consistency score is calculated based on the space error after affine transformation mapping, and the comprehensive score is the weighted sum of the template point and the flying point. Further, the fixed light field offset matrix is calibrated in advance based on the spatial tran