CN117191815A - A Mini/Micro LED small target defect detection method and classification system - Google Patents
A Mini/Micro LED small target defect detection method and classification system Download PDFInfo
- Publication number
- CN117191815A CN117191815A CN202311157181.4A CN202311157181A CN117191815A CN 117191815 A CN117191815 A CN 117191815A CN 202311157181 A CN202311157181 A CN 202311157181A CN 117191815 A CN117191815 A CN 117191815A
- Authority
- CN
- China
- Prior art keywords
- mini
- micro led
- chip
- defect
- led panel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
本发明公开了一种Mini/Micro LED小目标缺陷检测方法与分类系统，在对Mini/Micro LED的芯片位置进行定位之后，采用不同算法分别对芯片进行缺陷检测与分类，再将Mini/Micro LED的芯片位置填充为Mini/Micro LED面板的颜色，根据预设切片比例，对Mini/Micro LED面板图像进行分割，对分割图像进行全局阈值分割，得到分割图像上的面板缺陷，将分割图像进行拼接，得到整个Mini/Micro LED面板图像上的缺陷位置坐标，实现了在准确对芯片缺陷进行检测分类的同时避免了在面板缺陷检测时误将芯片误检成缺陷。
The invention discloses a Mini/Micro LED small target defect detection method and classification system. After locating the chip position of the Mini/Micro LED, different algorithms are used to detect and classify the defects of the chip respectively, and then the Mini/Micro LED The chip position is filled with the color of the Mini/Micro LED panel. According to the preset slicing ratio, the Mini/Micro LED panel image is segmented, and the segmented image is segmented with a global threshold. The panel defects on the segmented image are obtained, and the segmented images are spliced. , the defect position coordinates on the entire Mini/Micro LED panel image are obtained, which enables accurate detection and classification of chip defects and avoids mistakenly detecting chips as defects during panel defect detection.
Description
Technical Field
The invention relates to the technical field of small target defect detection, in particular to a Mini/Micro LED small target defect detection method and a classification system.
Background
With the rapid development of intelligent systems and the wide application of portable photographing devices, a large number of small targets exist in videos and images, and key information needs to be acquired from the small targets in many fields, which fully indicates that the detection of the small targets has great research value and application prospect. However, compared with large and medium-sized target detection, the performance index of the same detection algorithm used for small target detection is often lower, and the expected effect is difficult to achieve. This is because the resolution of a small target is low, the information is limited, key features are easily lost in the downsampling process, and the phenomena of missed detection and false detection are serious.
Compared with the conventional LED, the size of the Mini/Micro LED is further reduced, the manufacturing process is more complex, and the size of the LED chip adopted by the Mini/Micro LED is in the micron level. The Mini/Micro LED small target defect detection comprises panel defect detection and chip defect detection, wherein defects on the panel mostly appear in the forms of stains, hairs and the like, the existence of the chip can influence the panel defect detection, and the chip is easy to be erroneously detected as the defect. Defects on the chip are mostly caused by solder operation errors, the types and the shapes of the defects are various, and the defect classification accuracy is low. Therefore, how to avoid the influence of the chip on the defect detection of the Mini/Micro LED panel and accurately classify the defects of the chip is a technical problem to be solved by the technicians in the field.
Disclosure of Invention
The embodiment of the invention provides a Mini/Micro LED small target defect detection method and a classification system, which are used for solving the technical problems that chips are easy to be mistakenly detected to be defects and the accuracy of classifying the chips is low when the defects of a Mini/Micro LED panel are detected.
In view of the above, the first aspect of the present invention provides a method for detecting small target defects of Mini/Micro LEDs, comprising:
acquiring the chip position of a Mini/Micro LED;
performing defect detection and classification on the chip by adopting different algorithms, wherein the chip defect types comprise cold joint, tin frying, leakage fixation, reflection fixation, offset, deflection and polarity inversion;
filling the chip positions of the Mini/Micro LEDs into the color of a Mini/Micro LED panel;
dividing the Mini/Micro LED panel image according to a preset slice proportion, and performing global threshold segmentation on the divided image to obtain panel defects on the divided image;
and splicing the segmented images to obtain the defect position coordinates on the whole Mini/Micro LED panel image.
Optionally, performing defect detection and classification on the chip by using different algorithms respectively includes:
detecting whether the chip has defects of cold joint, tin frying and leakage fixation by using a Yolov5 detection model;
calculating a polarity characteristic vector of the chip, and judging whether the chip has a solid-state defect according to the polarity characteristic vector, wherein if the polarity characteristic vector is 0, the chip is judged to have the solid-state defect;
calculating a distance average value of azimuth feature vectors of the chip and the bonding pad, and judging whether the chip has an offset defect according to the distance average value, wherein if the distance average value is larger than a preset threshold value, the chip is judged to have the offset defect;
calculating the average deviation of the deflection angles of the polar characteristic vector and the (1, 0) vector of the chip, and judging whether the chip has deflection and polarity reversal defects according to the average deviation of the deflection angles, wherein if the average deviation of the deflection angles is larger than a preset angle threshold value, the chip is judged to have deflection defects, and if the average deviation of the deflection angles is larger than 90 degrees, the chip is judged to have polarity reversal defects, and the preset angle threshold value is not larger than 90 degrees.
Optionally, acquiring the chip position of the Mini/Micro LED further comprises:
acquiring a Mini/Micro LED panel image;
and (5) performing position correction on the Mini/Micro LED panel image.
Optionally, acquiring the Mini/Micro LED panel image includes:
and acquiring images of the whole Mini/Micro LED panel according to a Z-shaped running route by using the left upper corner of the Mini/Micro LED panel as a starting position through a camera, and splicing all the images to obtain the images of the Mini/Micro LED panel.
Optionally, performing position correction on the Mini/Micro LED panel image includes:
acquiring a panel upper left reference Mask point coordinate P1 and a panel lower right reference Mask point coordinate P2 in a Mini/Micro LED panel image, and acquiring an upper left reference Mask point coordinate P3 and a panel lower right reference Mask point coordinate P4 in a standard Mini/Micro LED panel;
calculating an included angle formed by P1P2 and P3P4 to obtain a homogeneous transformation matrix;
and carrying out position correction on the Mini/Micro LED panel image according to the homogeneous transformation matrix.
The second aspect of the invention provides a Mini/Micro LED small target defect classification system, comprising:
the chip positioning module is used for acquiring the chip position of the Mini/Micro LED;
the chip defect classification module is used for respectively carrying out defect detection and classification on the chip by adopting different algorithms, wherein the chip defect types comprise cold joint, tin frying, solid leakage, solid inversion, offset, deflection and polarity inversion;
the panel filling module is used for filling the chip positions of the Mini/Micro LEDs into the color of the Mini/Micro LED panel;
the segmentation module is used for segmenting the Mini/Micro LED panel image according to a preset slicing proportion, and performing global threshold segmentation on the segmented image to obtain panel defects on the segmented image;
and the panel defect positioning module is used for splicing the segmented images to obtain the defect position coordinates of the whole Mini/Micro LED panel image.
Optionally, the chip defect classification module is specifically configured to:
detecting whether the chip has defects of cold joint, tin frying and leakage fixation by using a Yolov5 detection model;
calculating a polarity characteristic vector of the chip, and judging whether the chip has a solid-state defect according to the polarity characteristic vector, wherein if the polarity characteristic vector is 0, the chip is judged to have the solid-state defect;
calculating a distance average value of azimuth feature vectors of the chip and the bonding pad, and judging whether the chip has an offset defect according to the distance average value, wherein if the distance average value is larger than a preset threshold value, the chip is judged to have the offset defect;
calculating the average deviation of the deflection angles of the polar characteristic vector and the (1, 0) vector of the chip, and judging whether the chip has deflection and polarity reversal defects according to the average deviation of the deflection angles, wherein if the average deviation of the deflection angles is larger than a preset angle threshold value, the chip is judged to have deflection defects, and if the average deviation of the deflection angles is larger than 90 degrees, the chip is judged to have polarity reversal defects, and the preset angle threshold value is not larger than 90 degrees.
Optionally, a correction module is also included;
the correction module is used for:
acquiring a Mini/Micro LED panel image;
and (5) performing position correction on the Mini/Micro LED panel image.
Optionally, acquiring the Mini/Micro LED panel image includes:
and acquiring images of the whole Mini/Micro LED panel according to a Z-shaped running route by using the left upper corner of the Mini/Micro LED panel as a starting position through a camera, and splicing all the images to obtain the images of the Mini/Micro LED panel.
Optionally, acquiring the Mini/Micro LED panel image includes:
and acquiring images of the whole Mini/Micro LED panel according to a Z-shaped running route by using the left upper corner of the Mini/Micro LED panel as a starting position through a camera, and splicing all the images to obtain the images of the Mini/Micro LED panel.
Optionally, performing position correction on the Mini/Micro LED panel image includes:
acquiring a panel upper left reference Mask point coordinate P1 and a panel lower right reference Mask point coordinate P2 in a Mini/Micro LED panel image, and acquiring an upper left reference Mask point coordinate P3 and a panel lower right reference Mask point coordinate P4 in a standard Mini/Micro LED panel;
calculating an included angle formed by P1P2 and P3P4 to obtain a homogeneous transformation matrix;
and carrying out position correction on the Mini/Micro LED panel image according to the homogeneous transformation matrix.
From the technical scheme, the Mini/Micro LED small target defect detection method provided by the invention has the following advantages:
according to the method for detecting the small target defects of the Mini/Micro LED, after the chip positions of the Mini/Micro LED are positioned, the chips are respectively detected and classified by adopting different algorithms, the chip positions of the Mini/Micro LED are filled with the colors of the Mini/Micro LED panel, the Mini/Micro LED panel image is segmented according to the preset slicing proportion, global threshold segmentation is carried out on the segmented image, the panel defects on the segmented image are obtained, the segmented image is spliced, and the defect position coordinates on the whole Mini/Micro LED panel image are obtained, so that the defect that the chips are mistakenly detected as defects when the panel defects are detected is avoided when the chip defects are accurately detected and classified, and the technical problems that the chips are mistakenly detected as the defects when the chip defects of the Mini/Micro LED panel are detected and the chip defect classification accuracy is lower are solved.
Drawings
For a clearer description of embodiments of the invention or of solutions according to the prior art, the figures which are used in the description of the embodiments or of the prior art will be briefly described, it being obvious that the figures in the description below are only some embodiments of the invention, from which, without the aid of inventive efforts, other relevant figures can be obtained for a person skilled in the art.
FIG. 1 is a schematic flow chart of a Mini/Micro LED small target defect detection method provided by the invention;
FIG. 2 is a schematic diagram of a standard Mini/Micro LED panel image (left) and a Mini/Micro LED panel image (right) to be corrected;
fig. 3 is a schematic structural diagram of the small target defect classification system of Mini/Micro LED provided by the invention.
Detailed Description
In order to make the present invention better understood by those skilled in the art, the following description will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
For ease of understanding, referring to fig. 1, an embodiment of a method for detecting small target defects of Mini/Micro LEDs is provided in the present invention, including:
and step 101, acquiring the chip position of the Mini/Micro LED.
After the Mini/Micro LED panel image is obtained, the chip position of the Mini/Micro LED is positioned by a preset positioning algorithm, wherein the chip positioning can be performed by a template matching method with a high detection speed.
In one embodiment, because the resolution of the camera is higher, the field of view of the camera is small, and the whole Mini/Micro LED panel image cannot be acquired at one time, the camera is enabled to acquire images according to a Z-shaped running route by taking the upper left corner of the Mini/Micro LED panel as a starting position and the lower right corner of the Mini/Micro LED panel as an ending position, and then all the images are spliced to obtain the whole Mini/Micro LED panel image. In order to avoid affecting the accuracy of subsequent processing results, the invention also needs to correct the position of the Mini/Micro LED panel image, specifically, as shown in fig. 2, obtain the top left reference Mask point coordinate P1 (X1, Y1) and the bottom right reference Mask point coordinate P2 (X2, Y2) of the panel in the Mini/Micro LED panel image, and obtain the top left reference Mask point coordinate P3 (X3, Y3) and the bottom right reference Mask point coordinate P4 (X4, Y4) in the standard Mini/Micro LED panel, and calculate the included angle θ formed by P1P2 and P3P4 to obtain a homogeneous transformation matrix. The calculation formula of the included angle theta is as follows:
θ＝tan -1 (α-β)
where α is the slope of the straight line P1P2, β is the slope of the straight line P3P4, θ is positive and indicates clockwise rotation, and θ is negative and indicates counterclockwise rotation.
The calculation formula of the translation amount is as follows:
where Δx is the x-axis translation and Δy is the y-axis translation.
The homogeneous transformation matrix includes a clockwise rotation matrix R1 and a counterclockwise rotation matrix R2:
and (3) carrying out position correction on the Mini/Micro LED panel image according to the homogeneous transformation matrix, namely enabling the Mini/Micro LED panel image to translate and rotate to be in position through the following steps:
(x y 1)＝(m n 1)·R1
or (b)
(x y 1)＝(m n 1)·R2
Where (x, y) is the transformed coordinates and (m, n) is the coordinates before transformation.
And 102, respectively detecting and classifying the defects of the chip by adopting different algorithms, wherein the types of the defects of the chip comprise cold joint, tin explosion, solid leakage, solid inversion, offset, deflection and polarity inversion.
The chip defect types include dummy solder, solder bumps, leakage, reverse, offset, deflection, and polarity reversal.
And (5) detecting the defects of cold joint, tin explosion and leakage fixation by using a Yolov5 detection model. The defect detection by using the Yolov5 detection model is the prior art, and will not be described in detail herein.
And calculating the polarity characteristic vector of the chip for the solid-inverse defect, and judging whether the chip has the solid-inverse defect according to the polarity characteristic vector, wherein if the polarity characteristic vector is judged to be 0, the chip is judged to have the solid-inverse defect. For polar components, the polarity direction of the component is generally identified on the component. Wherein, a circular mark is arranged near a certain section of the LED chip and used as a polarity mark. The identified regions may be extracted using conventional image processing operators. Finally, from the extracted chip center coordinates P 0 Point to polar region center coordinate P p The polar feature vector P is obtained 0 P p 。
And calculating a distance average value d of azimuth feature vectors of the chip and the bonding pad for the offset defect, and judging whether the chip has the offset defect according to the distance average value d, wherein if the distance average value d is larger than a preset threshold delta, the chip is judged to have the offset defect. After the coarse positioning of the chip is obtained according to the template matching algorithm, the precise positioning of the identification chip can be obtained by adopting a minimum circumscribed rectangle method. The azimuth feature vector comprises the absolute coordinates P of the 4 corner points and the center point of the minimum circumscribed rectangle on the substrate 1 、P 2 、P 3 、P 4 、P 0 . Finally, the azimuth feature vector of the identification target is marked as V= { P 1 ,P 2 ,P 3 ,P 4 ,P 0 }。
The calculation formula of the distance average d is:
wherein,and->The abscissa and the ordinate of the center point of the chip azimuth feature vectorCoordinates of->And->The abscissa and the ordinate of the center point of the pad azimuth feature vector, respectively.
For deflection and polarity reversal defects, the polarity eigenvector T and the (1, 0) vector T of the chip are calculated 0 Average difference of deflection angles of (2)According to the mean deviation of deflection angles->Judging whether the chip has deflection and polarity reversal defects, wherein if the deflection angle is average difference +.>If the deflection angle is larger than the preset angle threshold value, judging that the chip has deflection defects, and if the deflection angle is equal to or smaller than the preset angle threshold value, judging that the chip has deflection defects>If the angle is larger than 90 degrees, judging that the chip has polarity reversal defects, and the preset angle threshold is not larger than 90 degrees.
Average difference of deflection anglesThe calculation formula of (2) is as follows:
and step 103, filling the chip positions of the Mini/Micro LEDs into the color of the Mini/Micro LED panel.
It should be noted that, since the chip may have an effect on the detection of the panel defect, the chip on the panel is first removed and the panel is filled with the same color. And extracting RGB values of a part of a panel around the chip, performing median calculation on a certain amount of RGB data, and filling the whole chip position according to the median RGB values serving as filling colors.
And 104, dividing the Mini/Micro LED panel image according to a preset slice proportion, and performing global threshold segmentation on the divided image to obtain panel defects on the divided image.
It should be noted that there may be small target defects such as dirt, hair, scratches, etc. on the panel, which are generally smaller than 50 μm, and have a large difference from the characteristics of the panel, and the panel may be locally thresholded to find out defects on the panel in consideration of the problems such as uneven lighting and shadows on the panel. Specifically, the image is segmented according to a certain slice proportion, and global threshold segmentation is carried out on the segmented image to obtain defects on the panel in each segmented image.
And 105, splicing the segmented images to obtain the defect position coordinates on the whole Mini/Micro LED panel image.
It should be noted that, after all the divided images are spliced together and the defective pixel points on the divided images are transformed, the defective coordinates on the complete image can be obtained, and the defective pixel point transformation formula is as follows:
wherein, (P' x ,P′ y ) Sitting for pixels on a complete imageLabel (P) x ,P y ) For pixel coordinates on slice images, n i Representing the i-th segmented image. n is n x Represents the maximum value of the transverse slice, n y The maximum value of the longitudinal slice is represented, (W, H) the size of the complete panel image, and (W, H) the size of the divided image.
When (when)Or->When it cannot be divided, n x Or n y The number is increased by one, and when the images are spliced, the abscissa of the last divided image in each row is (W-W), the ordinate of the last divided image in each column is (H-H), and the pixel coordinates of the divided images are added on the basis.
According to the method for detecting the small target defects of the Mini/Micro LED, after the chip positions of the Mini/Micro LED are positioned, the chips are respectively detected and classified by adopting different algorithms, the chip positions of the Mini/Micro LED are filled with the colors of the Mini/Micro LED panel, the Mini/Micro LED panel image is segmented according to the preset slicing proportion, global threshold segmentation is carried out on the segmented image, the panel defects on the segmented image are obtained, the segmented image is spliced, and the defect position coordinates on the whole Mini/Micro LED panel image are obtained, so that the defect that the chips are mistakenly detected as defects when the panel defects are detected is avoided when the chip defects are accurately detected and classified, and the technical problems that the chips are mistakenly detected as the defects when the chip defects of the Mini/Micro LED panel are detected and the chip defect classification accuracy is lower are solved.
For ease of understanding, referring to fig. 3, an embodiment of a Mini/Micro LED small target defect classification system is provided in the present invention, including:
the chip positioning module is used for acquiring the chip position of the Mini/Micro LED;
the chip defect classification module is used for respectively carrying out defect detection and classification on the chip by adopting different algorithms, wherein the chip defect types comprise cold joint, tin frying, solid leakage, solid inversion, offset, deflection and polarity inversion;
the panel filling module is used for filling the chip positions of the Mini/Micro LEDs into the color of the Mini/Micro LED panel;
the segmentation module is used for segmenting the Mini/Micro LED panel image according to a preset slicing proportion, and performing global threshold segmentation on the segmented image to obtain panel defects on the segmented image;
and the panel defect positioning module is used for splicing the segmented images to obtain the defect position coordinates of the whole Mini/Micro LED panel image.
The chip defect classification module is specifically used for:
detecting whether the chip has defects of cold joint, tin frying and leakage fixation by using a Yolov5 detection model;
calculating a polarity characteristic vector of the chip, and judging whether the chip has a solid-state defect according to the polarity characteristic vector, wherein if the polarity characteristic vector is 0, the chip is judged to have the solid-state defect;
calculating a distance average value of azimuth feature vectors of the chip and the bonding pad, and judging whether the chip has an offset defect according to the distance average value, wherein if the distance average value is larger than a preset threshold value, the chip is judged to have the offset defect;
calculating the average deviation of the deflection angles of the polar characteristic vector and the (1, 0) vector of the chip, and judging whether the chip has deflection and polarity reversal defects according to the average deviation of the deflection angles, wherein if the average deviation of the deflection angles is larger than a preset angle threshold value, the chip is judged to have deflection defects, and if the average deviation of the deflection angles is larger than 90 degrees, the chip is judged to have polarity reversal defects, and the preset angle threshold value is not larger than 90 degrees.
The system also comprises a correction module;
the correction module is used for:
acquiring a Mini/Micro LED panel image;
and (5) performing position correction on the Mini/Micro LED panel image.
Acquiring a Mini/Micro LED panel image, comprising:
and acquiring images of the whole Mini/Micro LED panel according to a Z-shaped running route by using the left upper corner of the Mini/Micro LED panel as a starting position through a camera, and splicing all the images to obtain the images of the Mini/Micro LED panel.
Performing position correction on the Mini/Micro LED panel image, including:
acquiring a panel upper left reference Mask point coordinate P1 and a panel lower right reference Mask point coordinate P2 in a Mini/Micro LED panel image, and acquiring an upper left reference Mask point coordinate P3 and a panel lower right reference Mask point coordinate P4 in a standard Mini/Micro LED panel;
calculating an included angle formed by P1P2 and P3P4 to obtain a homogeneous transformation matrix;
and carrying out position correction on the Mini/Micro LED panel image according to the homogeneous transformation matrix.
The Mini/Micro LED small target defect classification system provided by the invention is used for executing the Mini/Micro LED small target defect detection method provided by the invention, and the principle is the same as that of the Mini/Micro LED small target defect detection method provided by the invention, and the detailed description is omitted.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. The Mini/Micro LED small target defect detection method is characterized by comprising the following steps of:
acquiring the chip position of a Mini/Micro LED;
performing defect detection and classification on the chip by adopting different algorithms, wherein the chip defect types comprise cold joint, tin frying, leakage fixation, reflection fixation, offset, deflection and polarity inversion;
filling the chip positions of the Mini/Micro LEDs into the color of a Mini/Micro LED panel;
dividing the Mini/Micro LED panel image according to a preset slice proportion, and performing global threshold segmentation on the divided image to obtain panel defects on the divided image;
and splicing the segmented images to obtain the defect position coordinates on the whole Mini/Micro LED panel image.
2. The method for detecting small target defects of Mini/Micro LED according to claim 1, wherein the method for detecting and classifying the defects of the chips by using different algorithms comprises:
detecting whether the chip has defects of cold joint, tin frying and leakage fixation by using a Yolov5 detection model;
calculating a polarity characteristic vector of the chip, and judging whether the chip has a solid-state defect according to the polarity characteristic vector, wherein if the polarity characteristic vector is 0, the chip is judged to have the solid-state defect;
calculating a distance average value of azimuth feature vectors of the chip and the bonding pad, and judging whether the chip has an offset defect according to the distance average value, wherein if the distance average value is larger than a preset threshold value, the chip is judged to have the offset defect;
calculating the average deviation of the deflection angles of the polar characteristic vector and the (1, 0) vector of the chip, and judging whether the chip has deflection and polarity reversal defects according to the average deviation of the deflection angles, wherein if the average deviation of the deflection angles is larger than a preset angle threshold value, the chip is judged to have deflection defects, and if the average deviation of the deflection angles is larger than 90 degrees, the chip is judged to have polarity reversal defects, and the preset angle threshold value is not larger than 90 degrees.
3. The method for detecting small target defects of Mini/Micro LEDs according to claim 1, wherein obtaining chip positions of Mini/Micro LEDs further comprises:
acquiring a Mini/Micro LED panel image;
and (5) performing position correction on the Mini/Micro LED panel image.
4. The Mini/Micro LED small target defect detection method according to claim 3, wherein obtaining the Mini/Micro LED panel image comprises:
and acquiring images of the whole Mini/Micro LED panel according to a Z-shaped running route by using the left upper corner of the Mini/Micro LED panel as a starting position through a camera, and splicing all the images to obtain the images of the Mini/Micro LED panel.
5. The Mini/Micro LED small target defect detection method according to claim 3, wherein performing position correction on the Mini/Micro LED panel image comprises:
acquiring a panel upper left reference Mask point coordinate P1 and a panel lower right reference Mask point coordinate P2 in a Mini/Micro LED panel image, and acquiring an upper left reference Mask point coordinate P3 and a panel lower right reference Mask point coordinate P4 in a standard Mini/Micro LED panel;
calculating an included angle formed by P1P2 and P3P4 to obtain a homogeneous transformation matrix;
and carrying out position correction on the Mini/Micro LED panel image according to the homogeneous transformation matrix.
6. The Mini/Micro LED small target defect classification system is characterized by comprising:
the chip positioning module is used for acquiring the chip position of the Mini/Micro LED;
the chip defect classification module is used for respectively carrying out defect detection and classification on the chip by adopting different algorithms, wherein the chip defect types comprise cold joint, tin frying, solid leakage, solid inversion, offset, deflection and polarity inversion;
the panel filling module is used for filling the chip positions of the Mini/Micro LEDs into the color of the Mini/Micro LED panel;
the segmentation module is used for segmenting the Mini/Micro LED panel image according to a preset slicing proportion, and performing global threshold segmentation on the segmented image to obtain panel defects on the segmented image;
and the panel defect positioning module is used for splicing the segmented images to obtain the defect position coordinates of the whole Mini/Micro LED panel image.
7. The Mini/Micro LED small target defect classification system of claim 6, wherein the chip defect classification module is specifically configured to:
detecting whether the chip has defects of cold joint, tin frying and leakage fixation by using a Yolov5 detection model;
calculating a polarity characteristic vector of the chip, and judging whether the chip has a solid-state defect according to the polarity characteristic vector, wherein if the polarity characteristic vector is 0, the chip is judged to have the solid-state defect;
calculating a distance average value of azimuth feature vectors of the chip and the bonding pad, and judging whether the chip has an offset defect according to the distance average value, wherein if the distance average value is larger than a preset threshold value, the chip is judged to have the offset defect;
calculating the average deviation of the deflection angles of the polar characteristic vector and the (1, 0) vector of the chip, and judging whether the chip has deflection and polarity reversal defects according to the average deviation of the deflection angles, wherein if the average deviation of the deflection angles is larger than a preset angle threshold value, the chip is judged to have deflection defects, and if the average deviation of the deflection angles is larger than 90 degrees, the chip is judged to have polarity reversal defects, and the preset angle threshold value is not larger than 90 degrees.
8. The Mini/Micro LED small target defect classification system of claim 6, further comprising a correction module;
the correction module is used for:
acquiring a Mini/Micro LED panel image;
and (5) performing position correction on the Mini/Micro LED panel image.
9. The Mini/Micro LED small target defect classification system of claim 8, wherein acquiring Mini/Micro LED panel images comprises:
and acquiring images of the whole Mini/Micro LED panel according to a Z-shaped running route by using the left upper corner of the Mini/Micro LED panel as a starting position through a camera, and splicing all the images to obtain the images of the Mini/Micro LED panel.
10. The Mini/Micro LED small target defect classification system of claim 8, wherein performing position correction on the Mini/Micro LED panel image comprises:
acquiring a panel upper left reference Mask point coordinate P1 and a panel lower right reference Mask point coordinate P2 in a Mini/Micro LED panel image, and acquiring an upper left reference Mask point coordinate P3 and a panel lower right reference Mask point coordinate P4 in a standard Mini/Micro LED panel;
calculating an included angle formed by P1P2 and P3P4 to obtain a homogeneous transformation matrix;
and carrying out position correction on the Mini/Micro LED panel image according to the homogeneous transformation matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311157181.4A CN117191815A (en) | 2023-09-08 | 2023-09-08 | A Mini/Micro LED small target defect detection method and classification system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311157181.4A CN117191815A (en) | 2023-09-08 | 2023-09-08 | A Mini/Micro LED small target defect detection method and classification system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117191815A true CN117191815A (en) | 2023-12-08 |
Family
ID=88990049
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311157181.4A Pending CN117191815A (en) | 2023-09-08 | 2023-09-08 | A Mini/Micro LED small target defect detection method and classification system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117191815A (en) |
-
2023
- 2023-09-08 CN CN202311157181.4A patent/CN117191815A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115791822B (en) | Visual inspection algorithm and inspection system for wafer surface defects | |
| US11227381B2 (en) | Substrate defect inspection apparatus, substrate defect inspection method, and storage medium | |
| WO2019237976A1 (en) | Differential image-based foreign matter detection method and apparatus, and device and storage medium | |
| CN109932370A (en) | System and method for white speckle detection with improved preprocessing | |
| CN109767428A (en) | A kind of camera module contamination detection method | |
| CN115082314B (en) | Method for splicing optical surface defect images step by adopting self-adaptive feature extraction | |
| CN112419225A (en) | SOP type chip detection method and system based on pin segmentation | |
| CN115527049A (en) | A high-precision measurement method for lead frame pin spacing | |
| CN111563883A (en) | Screen visual positioning method, positioning device and storage medium | |
| Pahwa et al. | Automated void detection in TSVs from 2D X-ray scans using supervised learning with 3D X-ray scans | |
| CN113658039A (en) | A method for determining the splicing sequence of medicine bottle label images | |
| CN117557531A (en) | Robust PIN pin alignment detection method, device and storage medium with sub-pixel accuracy | |
| CN113643245B (en) | Method and device for measuring screen defects and computer readable storage medium | |
| CN110874837A (en) | Automatic defect detection method based on local feature distribution | |
| CN111199563B (en) | Single crystalline silicon solar wafer image positioning method based on improved Hough transform algorithm | |
| CN117191815A (en) | A Mini/Micro LED small target defect detection method and classification system | |
| Weng et al. | Development of an adaptive template for fast detection of lithographic patterns of light-emitting diode chips | |
| CN119850511A (en) | Component offset detection method and device and electronic equipment | |
| CN111462242B (en) | A Rectangular Pin Chip Positioning Method Based on Improved Deformable Part Model | |
| CN116843659A (en) | An automatic detection method for circuit board faults based on infrared images | |
| Deng et al. | Learned optical flow-based defect detection with coarse-to-fine nonrigid image alignment | |
| CN120997538B (en) | A method and apparatus for identifying chip unpacking | |
| CN114298994A (en) | Positioning method and device based on Gaussian filtering and utilizing chip key points | |
| CN114187294B (en) | Regular wafer positioning method based on prior information | |
| CN120375105B (en) | A method for defect classification and detection based on sub-pixel template matching |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |