CN115860145A - Machine learning system and machine learning method for chip position correctness - Google Patents
Machine learning system and machine learning method for chip position correctness Download PDFInfo
- Publication number
- CN115860145A CN115860145A CN202211677411.5A CN202211677411A CN115860145A CN 115860145 A CN115860145 A CN 115860145A CN 202211677411 A CN202211677411 A CN 202211677411A CN 115860145 A CN115860145 A CN 115860145A
- Authority
- CN
- China
- Prior art keywords
- image data
- chip
- training image
- base
- data
- 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
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a machine learning system and a machine learning method for chip position correctness, wherein the machine learning method comprises the following steps: acquiring chip image data and base image data in the training image data according to the brightness characteristics of the chip and the base; judging whether the position deviation phenomenon occurs or not according to the relative position relation between the chip image data and the base image data; when the position deviation phenomenon occurs, marking the training image data as unqualified data; and when the position deviation phenomenon does not occur, marking the training image data as qualified data.
Description
Technical Field
The present invention relates to a system and a method for training the position accuracy of an electronic component, and more particularly, to a system and a method for machine learning the position accuracy of a chip.
Background
The existing chip positioning equipment firstly places a chip on a base through a robot arm, and then sequentially emits laser to two opposite angle positions of the chip by using an optical distance measuring instrument. When the optical distance meter receives the reflected light from the two opposite corners, the distance between the two opposite corners of the chip and the optical distance meter can be estimated according to the time difference between the time point of emitting the laser and the time point of receiving the reflected light.
When the distance between the first opposite angle of the chip and the optical distance meter is equal to the distance between the second opposite angle of the chip and the optical distance meter, the chip is placed correctly. On the contrary, when the distance between the first opposite angle of the chip and the optical distance meter is not equal to the distance between the second opposite angle of the chip and the optical distance meter, the chip is placed incorrectly. And it takes some time each time a distance measurement is made through the optical distance meter.
Disclosure of Invention
The technical problem to be solved by the present invention is to provide a machine learning system for chip position correctness, aiming at the deficiency of the prior art, comprising: a chip positioning device; the server is electrically connected to the chip positioning equipment, acquires a plurality of training image data from the chip positioning equipment, and executes the following operations on each training image data: distinguishing chip image data and base image data from training image data according to the chip brightness characteristics and the base brightness characteristics; judging whether the position deviation phenomenon occurs or not according to the relative position relation between the chip image data and the base image data; when the position deviation phenomenon occurs, marking the training image data as unqualified data; when the position deviation phenomenon does not occur, marking the training image data as qualified data; after the plurality of training image data are marked, the server trains the plurality of training image data by using a machine learning algorithm so as to generate a machine learning model.
Preferably, the server determining whether the position deviation phenomenon occurs includes: performing principal component analysis on the chip image data to obtain a chip central point; performing principal component analysis on the base image data to obtain a base center point; calculating an offset distance between the center point of the chip and the center point of the base; calculating an offset angle between the chip and the base; normalizing the offset distance to obtain an offset distance ratio; when the offset distance ratio is larger than a critical distance ratio or the offset angle is larger than a critical angle, marking the training image data as the unqualified data; and when the offset distance ratio is smaller than or equal to the critical distance ratio and the offset angle is smaller than or equal to the critical angle, marking the training image data as the qualified data.
Preferably, the machine learning algorithm is a support vector machine, and the machine learning model is a support vector machine model, the support vector machine model comprises a data clustering boundary, and two opposite sides of the data clustering boundary are respectively a position-qualified region and a position-unqualified region.
Preferably, the server further performs the following operations on each piece of training image data: establishing a frame line surrounding the chip image data; obtaining a first area surrounded by the frame line and a second area of the chip image data; obtaining the area difference between the first area and the second area; obtaining an area ratio of the area difference to the second area; when the area ratio is larger than a critical area ratio, judging that a plurality of chips are placed on the base; and when the area ratio is smaller than or equal to the critical area ratio, judging that only a single chip is placed on the base.
Preferably, the server further performs the following operations for each training image data: the server judges whether the training image data has a chip brightness characteristic or not; when the chip brightness feature does not exist in the training image data, marking the training image data as the unqualified data.
Preferably, the chip positioning device obtains the training image data and encodes the training image data into a plurality of string data, and when the server receives the plurality of string data, the server decodes the plurality of string data to obtain the plurality of training image data.
The technical problem to be solved by the present invention is to provide a machine learning method for chip position correctness, which is executed by a chip positioning device and a server, aiming at the deficiency of the prior art, and the machine learning method comprises: obtaining a plurality of training image data by the chip positioning equipment; obtaining, by the server, the plurality of training image data from the chip positioning device, and the server performing the following operations on each training image data: distinguishing a chip image data and a base image data from the training image data according to a chip brightness characteristic and a base brightness characteristic; judging whether a position offset phenomenon occurs according to the relative position relation between the chip image data and the base image data; when the position deviation phenomenon occurs, marking the training image data as unqualified data; and when the position deviation phenomenon does not occur, marking the training image data as qualified data; after the plurality of training image data are marked, the server trains the plurality of training image data by using a machine learning algorithm to generate a machine learning model.
Preferably, the server determining whether the position deviation phenomenon occurs includes: performing principal component analysis on the chip image data to obtain a chip central point; performing principal component analysis on the base image data to obtain a base center point; calculating an offset distance between the center point of the chip and the center point of the base; calculating an offset angle between the chip and the base; normalizing the offset distance to obtain an offset distance ratio; when the offset distance ratio is larger than a critical distance ratio or the offset angle is larger than a critical angle, marking the training image data as the unqualified data; and when the offset distance ratio is smaller than or equal to the critical distance ratio and the offset angle is smaller than or equal to the critical angle, marking the training image data as the qualified data.
Preferably, the server further performs the following operations on each training image data: establishing a frame line surrounding the chip image data; obtaining a first area surrounded by the frame line and a second area of the chip image data; obtaining the area difference between the first area and the second area; obtaining an area ratio of the area difference to the second area; when the area ratio is larger than a critical area ratio, judging that a plurality of chips are placed on the base; and when the area ratio is smaller than or equal to the critical area ratio, judging that only a single chip is placed on the base.
Preferably, the server further performs the following operations on each training image data: the server judges whether the training image data has a chip brightness characteristic or not; when the chip brightness feature does not exist in the training image data, marking the training image data as the unqualified data.
Preferably, the server determining whether the position deviation phenomenon occurs includes: performing principal component analysis on the chip image data to obtain a chip central point; performing principal component analysis on the base image data to obtain a base center point; calculating an offset distance between the center point of the chip and the center point of the base; calculating an offset angle between the chip and the base; judging whether the offset distance is larger than a critical distance; when the offset distance ratio is larger than the critical distance, marking the training image data as the unqualified data; when the offset distance is not greater than the critical distance, judging whether the offset angle is greater than a critical angle; when the deviation angle is larger than the critical angle, marking the training image data as the unqualified data; when the deviation angle is not larger than the critical angle, the training image data is marked as the qualified data.
The chip position accuracy machine learning system and the chip position accuracy machine learning method provided by the invention have the beneficial effects that the machine learning training is carried out according to the complete image data, so that the misjudgment probability of the chip position can be reduced. Compared with the prior art, after the chip is placed on the base by the chip positioning equipment, the distance between the two opposite angle positions of the chip is measured by the laser distance measuring instrument after the image is shot, and whether the placing position of the chip is correct or not can be confirmed. The machine learning system of the invention only needs to directly shoot and process images after the chip is arranged on the base, and the process is faster than laser ranging, thereby accelerating the processing speed.
For a better understanding of the features and technical content of the present invention, reference should be made to the following detailed description and accompanying drawings, which are provided for purposes of illustration and description only and are not intended to limit the invention.
Drawings
FIG. 1 is a functional block diagram of a chip location correctness machine learning system according to the present invention.
FIG. 2 is a flowchart illustrating a first embodiment of a method for machine learning chip location correctness according to the present invention.
Fig. 3 is a schematic diagram illustrating a chip image data and a base image data differentiated by a brightness characteristic.
FIG. 4 is a schematic diagram of obtaining a center point of a chip, a center point of a pedestal, and an offset angle between the chip and the pedestal through principal component analysis.
FIG. 5 is a diagram of an embodiment of a trained machine learning model of the present invention.
Fig. 6A and 6B are flowcharts illustrating a second embodiment of the method for machine learning chip location correctness according to the present invention.
FIG. 7 is a flowchart of a method for the server to determine whether a plurality of chips are stacked on the base.
FIG. 8 is a diagram illustrating an embodiment of creating a frame line enclosing chip image data.
FIG. 9 is a flowchart illustrating a third embodiment of the method for machine learning chip location correctness according to the present invention.
Detailed Description
The following is a description of the implementation of the "system and method for machine learning with chip location correctness" provided by the present invention with specific embodiments, and those skilled in the art can understand the advantages and effects of the present invention from the content provided in the present specification. The invention is capable of other and different embodiments and its several details are capable of modifications and various changes in detail without departing from the spirit and scope of the present invention. The drawings of the present invention are for illustrative purposes only and are not intended to be drawn to scale. The following embodiments will further explain the related art of the present invention in detail, but the contents are not provided to limit the scope of the present invention.
It will be understood that, although the terms first, second, third, etc. may be used herein to describe various components or signals, these components or signals should not be limited by these terms. These terms are used primarily to distinguish one element from another element or from one signal to another signal. In addition, the term "or" as used herein should be taken to include any one or combination of more of the associated listed items as the case may be.
FIG. 1 is a functional block diagram of a chip location correctness machine learning system according to the present invention. As shown in fig. 1, the machine learning system for chip position correctness includes a chip positioning device 1 and a server 2, and the server 2 is electrically connected to the chip positioning device 1. The chip positioning apparatus 1 includes a camera 11, and the camera 11 shoots towards the base to obtain the training image data. When the chip is placed on the base, the training image data comprises chip image data and base image data. The chip positioning device 1 transmits the training image data to the server 2. When the server 2 receives the training image data, the server 2 distinguishes the chip image data and the base image data from the training image data, and then the server 2 judges whether the position deviation phenomenon occurs according to the relative position relationship between the chip image data and the base image data. When the position deviation occurs, the server 2 marks the training image data as the non-qualified data. When the position deviation phenomenon does not occur, the server 2 marks the training image data as qualified data.
FIG. 2 is a flowchart illustrating a first embodiment of a method for machine learning chip location correctness according to the present invention. As shown in fig. 2, in step S201, the camera 11 of the chip positioning apparatus 1 acquires training image data. In step S203, the server 2 acquires training image data from the chip positioning apparatus 1. In step S205, the server 2 distinguishes the chip image data and the base image data from the training image data according to the chip brightness feature and the base brightness feature.
Fig. 3 is a schematic diagram of an embodiment of the server 2 distinguishing the chip image data and the base image data from the training image data. As shown in fig. 3, when the chip is placed on the base, the camera 11 shoots the base to obtain training image data, wherein the chip brightness characteristic of the chip is that a plurality of data points with brightness values higher than the brightness critical value are concentrated in a rectangular region, the base brightness characteristic of the base is that a plurality of data points with brightness values higher than the brightness critical value are distributed in four arc regions C1-C4, wherein the arc region C1 and the arc region C2 are located on a diagonal line, the arc region C3 and the arc region C4 are located on another diagonal line, and the brightness values of the region surrounded by the arc regions C1-C4 are lower than the brightness critical value. The server 2 distinguishes the chip image data M11 and the base image data M12 from the training image data M1 according to the chip luminance characteristics and the base luminance characteristics.
In step S207, the server 2 performs Principal Component Analysis (PCA) on the chip image data to obtain a chip center point. In step S209, the server 2 performs principal component analysis on the base image data to obtain the base center point.
FIG. 4 is a diagram illustrating an embodiment of the server 2 using principal component analysis to obtain the center point of the chip center and the center point of the base and the offset angle between the chip and the base. As shown in fig. 4, the server 2 performs principal component analysis on the chip image data M11 to obtain a chip center point P1 and a first major axis L1 and a first minor axis W1 passing through the chip center point P1, and the two-dimensional coordinates of the chip center point P1 are (X1, Y1). Similarly, the server 2 performs principal component analysis on the base image data M12 to obtain a base center point P2, and a second major axis L2 and a second minor axis W2 passing through the base center point P2, where the two-dimensional coordinates of the base center point P2 are (X2, Y2).
Regarding step S211, the server 2 calculates an offset distance between the chip center point P1 and the susceptor center point P2. For example, when the two-dimensional coordinates of the chip center point P1 are (X1, Y1) and the coordinates of the pad center point P2 are (X2, Y2), the offset distance is calculated by the formula
With respect to step S213, the server 2 calculates an offset angle between the chip and the susceptor. For example, the first major axis L1 is perpendicular to the first minor axis W1, and the second major axis L2 is perpendicular to the second minor axis W2. When the included angle between the first long axis L1 and the second long axis L2 is 10 degrees, and the included angle between the first short axis W1 and the second short axis W2 is also 10 degrees, the offset angle between the chip and the base is 10 degrees.
Regarding step S215, the server 2 normalizes the offset distance between the chip center point P1 and the pad center point P2 to obtain an offset distance ratio, wherein the normalized formula is (offset distance between the chip center point P1 and the pad center point P2)/(diagonal length of the chip), but the invention is not limited thereto. By the standardized processing, even if the resolution of the camera 12 changes, the offset distance ratio can accurately reflect the offset degree of the chip relative to the base.
With respect to step S217, the server 2 determines whether the offset distance ratio is greater than a preset threshold distance ratio. If yes, step S219 follows. If not, step S221 is followed. In step S219, the server 2 marks the training image data as the non-compliant data. With respect to step S221, the server 2 determines whether the offset angle is greater than the critical offset angle. If yes, step S223 follows. If not, step S225 follows.
In step S223, the server 2 marks the training image data as the non-compliant data. In step S225, the server 2 marks the training image data as the qualified data.
After steps S219, S223, and S225, step S227 follows. In step S227, the server 2 determines whether the amount of training image data reaches the amount threshold. If yes, step S229 follows. If not, the process returns to step S201.
With respect to step S229, the server 2 trains the plurality of marked training image data using a machine learning algorithm to generate a machine learning model.
FIG. 5 is a diagram of an embodiment of a machine learning model of the present invention. As shown in fig. 5, the server 2 uses a support vector machine (support vector machine) to train the plurality of marked training image data to generate a support vector machine model. The SVM model includes a data clustering boundary L, and two opposite sides of the data clustering boundary L are a PASS-location area PASS and a fail-location area NG, respectively. When the support vector machine model receives the test image data, the support vector machine model can judge that the test image data belongs to a position-qualified area PASS or a position-unqualified area NG. When the test image data belongs to the position qualified area PASS, the chip is correctly placed on the base. Otherwise, when the test image data belongs to the area with unqualified position, the chip is not correctly placed on the base. Whether the test result is qualified or unqualified, the server 2 will transmit the test result back to the chip positioning device 1, and the chip positioning device 1 can adjust the positioning parameters according to the test result of the support vector machine model.
Fig. 6A and 6B are flowcharts illustrating a second embodiment of the method for machine learning chip location correctness according to the present invention. As shown in fig. 6A, in step S601, the camera 11 of the chip positioning apparatus 1 acquires training image data. In step S603, the server 2 acquires training image data from the chip positioning apparatus 1. In step S605, the server 2 determines whether or not the chip is placed on the base, and if so, then step S607 is followed. If not, step S609 follows. With respect to step S607, the server 2 determines whether the number of chips placed on the base is single. In step S609, the server 2 marks the training image data as the non-compliant data.
How the server 2 determines whether the chip is placed on the base or not, specifically, if the server 2 detects the brightness characteristic of the chip in the training image data, it indicates that the chip is placed on the base. On the contrary, if the server 2 cannot detect the chip brightness feature in the training image data, it indicates that no chip is placed on the base.
If the server 2 determines that the number of chips placed on the susceptor is single, it follows step S611. If the server 2 determines that the number of chips placed on the susceptor is not single, step S613 is followed.
In step S611, the server 2 distinguishes the chip image data and the base image data from the training image data according to the chip brightness feature and the base brightness feature. In step S613, the server 2 marks the training image data as the non-compliant data.
After step S611, step S615 follows. In step S615, the server 2 performs a principal component analysis on the chip image data to obtain a chip center point. In step S617, the server 2 performs principal component analysis on the base image data to obtain a base center point.
As shown in fig. 6B, with respect to step S619, the server 2 calculates an offset distance between the chip center point P1 and the susceptor center point P2. With respect to step S621, the server 2 calculates an offset angle between the chip and the base.
Regarding step S623, the server 2 normalizes the offset distance between the chip center point P1 and the susceptor center point P2 to obtain an offset distance ratio.
With respect to step S625, the server 2 determines whether the offset distance ratio is greater than the critical distance ratio. If yes, step S627 follows. If not, step S629 follows. In step S627, the server 2 marks the training image data as failing data. With respect to step S629, the server 2 determines whether the offset angle is larger than the critical angle. If yes, step S631 follows. If not, step S633 follows.
In step S631, the server 2 marks the training image data as the unsatisfactory data. In step S633, the server 2 marks the training image data as the qualified data.
After steps S609, S613, S627, S631, and S633, step S635 follows. In step S635, the server 2 determines whether the amount of training image data reaches a threshold amount. If yes, step S637 follows. If not, the process returns to step S601.
Regarding step S637, the server 2 trains the multi-labeled training image data using a machine learning algorithm to generate a machine learning model.
Fig. 7 is a flowchart of a method for the server 2 to determine whether a plurality of chips are placed on the base, that is, fig. 7 further describes the sub-steps of step S607 of fig. 6. As shown in fig. 7, in step S701, the server 2 creates a frame line surrounding the chip image data. In step S703, the server 2 obtains a first area surrounded by the frame line and a second area of the chip image data. With respect to step S705, the server 2 calculates an area difference between the first area and the second area. With respect to step S707, the server 2 calculates an area ratio of the area difference to the second area. With respect to step S709, server 2 determines whether the area ratio is less than or equal to the critical area ratio. If yes, step S711 follows. If not, step S713 is followed. Regarding step S711, the server 2 determines that only one chip is placed on the cradle, followed by step S611. With respect to step S713, the server 2 determines that a plurality of chips are placed on the susceptor, followed by step S613.
Fig. 8 is a schematic diagram of an embodiment of creating a frame line enclosing chip image data, as shown in fig. 8, the server 2 creates a minimum frame line C enclosing the chip image data M11, and when the percentage of the area enclosed by the chip image data M11 compared to the frame line C is greater than a threshold percentage, it indicates that there is a great probability that only one chip is placed on the substrate. On the contrary, when the percentage of the area of the chip image data M11 to the area surrounded by the frame line C is smaller than the critical percentage, it means that there is a high probability that a plurality of chips are placed on the substrate.
FIG. 9 is a flowchart illustrating a third embodiment of the method for machine learning chip location correctness according to the present invention. The machine learning method of fig. 9 includes steps S901 to S927, wherein steps S901 to S913 are respectively the same as steps S201 to S213 of fig. 2, and steps S917 to S927 are respectively the same as steps S219 to S229 of fig. 2, with the difference being step S915. With respect to step S915, the server 2 determines whether the offset distance is greater than the critical distance. If so, step S917 follows. If not, step S919 follows.
Furthermore, regarding the data transmission between the chip positioning apparatus 1 and the server 2, for example, the chip positioning apparatus 1 may encode the training image data to form string data through the base 64 encoding rule. Then, the chip positioning device 1 transmits the string data to the server 2 through a Transmission Control Protocol (Transmission Control Protocol). When the server 2 receives the string data from the chip positioning device 1, the string data is decoded to obtain training image data, thereby solving the problem that the cross-platform object cannot be identified. The above data transmission methods are only examples, and the invention is not limited thereto.
Advantageous effects of the embodiment ]:
the machine learning system and the machine learning method for chip position correctness provided by the invention have the beneficial effects that the server performs machine learning training according to complete image data, so that the misjudgment probability of the chip position can be reduced. Compared with the prior art, after the chip is placed on the base by the chip positioning equipment, the distance measurement is carried out at two diagonal positions of the chip of the laser range finder except for the image shooting, so that whether the placing position of the chip is correct or not can be confirmed. The machine learning system of the invention only needs to directly shoot and process images after the chip is arranged on the base, and the process is faster than laser ranging, so the process speed is increased, the machine learning system of the chip position correctness and the machine learning method of the chip position correctness provided by the invention can reduce the misjudgment rate according to the evaluation model obtained by the complete image data training.
The above-mentioned embodiments are only preferred embodiments of the present invention, and not intended to limit the claims of the present invention, so that all the equivalent technical changes made by using the contents of the present specification and drawings are included in the claims of the present invention.
Claims (11)
1. A system for machine learning chip location correctness, comprising:
a chip positioning device;
a server electrically connected to the chip positioning device, the server obtaining a plurality of training image data from the chip positioning device, and the server performing the following operations on each training image data:
distinguishing a chip image data and a base image data from the training image data according to a chip brightness characteristic and a base brightness characteristic;
judging whether a position offset phenomenon occurs according to the relative position relation between the chip image data and the base image data;
when the position deviation phenomenon occurs, marking the training image data as unqualified data; and
when the position deviation phenomenon does not occur, marking the training image data as qualified data;
after the plurality of training image data are marked, the server trains the plurality of training image data by using a machine learning algorithm so as to generate a machine learning model.
2. The system of claim 1, wherein the server determining whether the shift phenomenon occurs comprises: performing principal component analysis on the chip image data to obtain a chip central point; performing principal component analysis on the base image data to obtain a base center point; calculating an offset distance between the center point of the chip and the center point of the base; calculating an offset angle between the chip and the base; normalizing the offset distance to obtain an offset distance ratio; when the offset distance ratio is larger than a critical distance ratio or the offset angle is larger than a critical angle, marking the training image data as the unqualified data; and when the offset distance ratio is smaller than or equal to the critical distance ratio and the offset angle is smaller than or equal to the critical angle, marking the training image data as the qualified data.
3. The system of claim 1, wherein the machine learning algorithm is SVM and the machine learning model is a SVM model, the SVM model includes a data boundary line, and the two opposite sides of the data boundary line are a qualified-location region and a unqualified-location region.
4. The system of claim 1, wherein the server further performs the following operations for each training image data: establishing a frame line surrounding the chip image data; obtaining a first area surrounded by the frame line and a second area of the chip image data; obtaining the area difference between the first area and the second area; obtaining an area ratio of the area difference to the second area; when the area ratio is larger than a critical area ratio, judging that a plurality of chips are placed on the base; and when the area ratio is smaller than or equal to the critical area ratio, judging that only a single chip is placed on the base.
5. The system of claim 1, wherein the server further performs the following operations for each training image data: the server judges whether the training image data has a chip brightness characteristic or not; when the chip brightness feature does not exist in the training image data, marking the training image data as the unqualified data.
6. The system of claim 1, wherein the chip positioning device obtains the training image data and encodes the training image data into a plurality of string data, and the server decodes the plurality of string data to obtain the plurality of training image data when receiving the plurality of string data.
7. A machine learning method for chip position correctness is executed by a chip positioning device and a server, and is characterized by comprising the following steps:
obtaining a plurality of training image data by the chip positioning equipment;
obtaining the plurality of training image data from the chip positioning device by the server, and executing the following operations on each training image data by the server:
distinguishing a chip image data and a base image data from the training image data according to a chip brightness characteristic and a base brightness characteristic;
judging whether a position offset phenomenon occurs according to the relative position relation between the chip image data and the base image data;
when the position deviation phenomenon occurs, marking the training image data as unqualified data; and
when the position deviation phenomenon does not occur, marking the training image data as qualified data;
after the plurality of training image data are marked, the server trains the plurality of training image data by using a machine learning algorithm to generate a machine learning model.
8. The method as claimed in claim 7, wherein the determining whether the position shift phenomenon occurs comprises: performing principal component analysis on the chip image data to obtain a chip central point; performing principal component analysis on the base image data to obtain a base center point; calculating an offset distance between the center point of the chip and the center point of the base; calculating an offset angle between the chip and the base; normalizing the offset distance to obtain an offset distance ratio; when the offset distance ratio is larger than a critical distance ratio or the offset angle is larger than a critical angle, marking the training image data as the unqualified data; and when the offset distance ratio is smaller than or equal to the critical distance ratio and the offset angle is smaller than or equal to the critical angle, marking the training image data as the qualified data.
9. The method as claimed in claim 7, wherein the server further performs the following operations for each training image data: establishing a frame line surrounding the chip image data; obtaining a first area surrounded by the frame line and a second area of the chip image data; obtaining the area difference between the first area and the second area; obtaining an area ratio of the area difference to the second area; when the area ratio is larger than a critical area ratio, judging that a plurality of chips are placed on the base; and when the area ratio is smaller than or equal to the critical area ratio, judging that only a single chip is placed on the base.
10. The method of claim 7, wherein the server further performs the following operations for each training image data: the server judges whether the training image data has a chip brightness characteristic or not; when the chip brightness characteristic does not exist in the training image data, marking the training image data as the unqualified data.
11. The method as claimed in claim 7, wherein the determining whether the position shift phenomenon occurs comprises: performing principal component analysis on the chip image data to obtain a chip central point; performing principal component analysis on the base image data to obtain a base center point; calculating an offset distance between the center point of the chip and the center point of the base; calculating an offset angle between the chip and the base; judging whether the offset distance is larger than a critical distance; when the offset distance ratio is larger than the critical distance, marking the training image data as the unqualified data; when the offset distance is not greater than the critical distance, judging whether the offset angle is greater than a critical angle; when the deviation angle is larger than the critical angle, marking the training image data as the unqualified data; when the deviation angle is not larger than the critical angle, marking the training image data as the qualified data.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211677411.5A CN115860145A (en) | 2022-12-26 | 2022-12-26 | Machine learning system and machine learning method for chip position correctness |
| TW112101668A TWI835524B (en) | 2022-12-26 | 2023-01-16 | Machine training system of chip position correctness and machine training method of chip position correctness |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211677411.5A CN115860145A (en) | 2022-12-26 | 2022-12-26 | Machine learning system and machine learning method for chip position correctness |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115860145A true CN115860145A (en) | 2023-03-28 |
Family
ID=85654872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211677411.5A Pending CN115860145A (en) | 2022-12-26 | 2022-12-26 | Machine learning system and machine learning method for chip position correctness |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN115860145A (en) |
| TW (1) | TWI835524B (en) |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101385125B (en) * | 2006-05-22 | 2011-04-13 | 株式会社尼康 | Exposure method and apparatus, maintenance method, and device manufacturing method |
| US9659204B2 (en) * | 2014-06-13 | 2017-05-23 | Conduent Business Services, Llc | Image processing methods and systems for barcode and/or product label recognition |
| JP6708217B2 (en) * | 2015-12-17 | 2020-06-10 | 株式会社ニコン | Pattern drawing device |
| EP3549099A4 (en) * | 2016-12-01 | 2020-07-29 | Berkeley Lights, Inc. | Automated detection and repositioning of micro-objects in microfluidic devices |
-
2022
- 2022-12-26 CN CN202211677411.5A patent/CN115860145A/en active Pending
-
2023
- 2023-01-16 TW TW112101668A patent/TWI835524B/en active
Also Published As
| Publication number | Publication date |
|---|---|
| TWI835524B (en) | 2024-03-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112115893B (en) | Instrument panel pointer reading identification method and device, computer equipment and storage medium | |
| US11002839B2 (en) | Method and apparatus for measuring angular resolution of multi-beam lidar | |
| US10964057B2 (en) | Information processing apparatus, method for controlling information processing apparatus, and storage medium | |
| CN106599897B (en) | Readings of pointer type meters recognition methods and device based on machine vision | |
| US9830491B2 (en) | Method and system for embedding and retrieving information through marker transformation | |
| US9787960B2 (en) | Image processing apparatus, image processing system, image processing method, and computer program | |
| WO2016092749A1 (en) | Information processing apparatus, information processing method, and program | |
| CN105717498B (en) | Laser range finder pitch angle measurement and correction system and method | |
| US20150053761A1 (en) | Portable encoded information reading terminal configured to acquire images | |
| CN111750804B (en) | Object measuring method and device | |
| US6519358B1 (en) | Parallax calculating apparatus, distance calculating apparatus, methods of the same, and information providing media | |
| CN113124763B (en) | Optical axis calibration method, device, terminal, system and medium for optical axis detection system | |
| CN108426527A (en) | A kind of RFID tag three-dimensional coordinate automatic testing method based on DLT | |
| CN105717499B (en) | Laser range finder deflects angular measurement and correction system and method | |
| CN116648727A (en) | Cross Spectral Feature Mapping for Camera Calibration | |
| CN111337910A (en) | Radar inspection method and device | |
| CN102706480A (en) | Automatic reading method for pointer thermometer | |
| US20170276701A1 (en) | Method, system for utilizing a probe card, and the probe card | |
| CN115860145A (en) | Machine learning system and machine learning method for chip position correctness | |
| EP4054187A1 (en) | Calibration method of a portable electronic device | |
| CN114882487A (en) | Pointer instrument reading method and system | |
| CN114170319A (en) | Method and device for adjusting test target | |
| US10451424B2 (en) | Position and azimuth detecting device | |
| US8120785B2 (en) | System and method for inspecting three-dimensional sensor | |
| CN110333185A (en) | Material detection method |
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 |