Vision-based chip placement system and operation method thereof

Abstract

The invention provides a vision-based chip placement system and an operation method thereof. The vision-based chip placement system comprises a driving module and a position detection module, the driving module is used for preliminarily placing a to-be-tested chip on the test board socket; the position detection module is used for detecting whether the chip to be tested is placed in place or not; moreover, if the to-be-tested chip is not placed in place, the position detection module is also used for calculating an offset parameter of the to-be-tested chip, and the driving module is also used for carrying out position adjustment on the to-be-tested chip based on the offset parameter, so that the to-be-tested chip is placed in place. According to the invention, the problem that the chip cannot be accurately placed in the preset position of the test board socket in the existing mechanical arm chip grabbing scheme can be effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic device processing, and more particularly, to a vision-based chip placement system and an operating method thereof. Background Art

[0002] In the field of electronic device processing, in order to ensure that the electrical performance of the processed chips is complete, the processed chips need to be electrically tested. During the test, the chip needs to be placed in the preset position of the test board socket first, and the chip is tested through the test board socket.

[0003] In the actual production process, in order to improve production efficiency, it is usually necessary to use a robotic arm to grab the chip and place the chip in the preset position of the test board socket. However, in the process of the robotic arm grabbing the chip and placing it in the test board socket, due to the displacement of the test board socket, the displacement of the chip tray and the movement error of the robotic arm itself, the chip is often placed in the incorrect position of the test board socket and cannot completely match the preset position. If the chip and the test board socket are directly contact tested at this time, the chip package is likely to be damaged due to the offset of the chip position; in addition, it is also possible to cause overload and short circuit due to incorrect pin contact, which in turn leads to the interruption of the test and causes economic losses.

[0004] Therefore, there is an urgent need for a solution that can accurately place the chip into the preset position of the test board socket. Summary of the Invention

[0005] In view of the above problems, the purpose of the present invention is to provide a vision-based chip placement system and its operation method to solve the problem that the existing robotic arm chip grabbing solution cannot accurately place the chip into the preset position of the test board socket.

[0006] The visual chip placement system provided by the present invention includes a driving module and a position detection module; wherein, The driving module is used to initially place the chip to be tested on the test board socket; The position detection module is used to detect whether the chip to be tested is placed in place; and If the chip to be tested is not placed in place, the position detection module is further used to calculate the offset parameter of the chip to be tested, and the driving module is further used to adjust the position of the chip to be tested based on the offset parameter to put the chip to be tested in place.

[0007] In addition, a preferred solution is that the driving module includes a mechanical arm, and the mechanical arm is used to grab the chip to be tested and shift the chip to be tested; and The position detection module includes a visual camera unit, which is arranged on the robotic arm and is used to capture visual data of the chip to be tested and the test board socket.

[0008] In addition, a preferred solution is that the position detection module further includes a target detection model, and the target detection model is used to determine whether the chip to be tested is placed in place based on the visual data of the chip to be tested and the visual data of the test board socket.

[0009] In addition, a preferred solution is that the position detection module further includes an offset detection model; and When the chip to be tested is not placed in place, the offset detection model is used to calculate the offset parameter of the chip to be tested based on the visual data of the chip to be tested and the visual data of the test board socket.

[0010] In addition, a preferred solution is that the target detection model and / or the offset detection model are constructed based on the YOLOv8 architecture.

[0011] In addition, a preferred solution is to further include a control module, wherein the control module is electrically connected to the driving module and the position detection module respectively; and The control module is used to send instruction signals to the driving module and the position detection module.

[0012] In another aspect, the present invention further provides a method for operating the aforementioned vision-based chip placement system, the method comprising: The chip to be tested is initially placed on the socket of the test board through the driving module; Detecting whether the chip to be tested is placed in place by a position detection module; If the chip to be tested is placed in place, the operation ends; If the chip under test is not in place, the position detection module calculates the offset parameter of the chip under test; and the driving module adjusts the position of the chip under test based on the offset parameter to put the chip under test in place, and then ends the operation.

[0013] In addition, a preferred solution is that the detecting whether the chip to be tested is placed in place by the position detection module includes: Detecting whether the visual data of the chip to be tested and the visual data of the test board socket meet a preset matching relationship; When the visual data of the chip to be tested and the visual data of the test board socket meet the preset matching relationship, it is determined that the chip to be tested is placed in place; when the visual data of the chip to be tested and the visual data of the test board socket do not meet the preset matching relationship, it is determined that the chip to be tested is not placed in place.

[0014] In addition, a preferred solution is that the calculating the offset parameter of the chip to be tested by the position detection module includes: Determining the geometric center position of the test board socket by visual data of the test board socket, and determining the geometric center position of the chip to be tested by visual data of the chip to be tested; The offset parameter is calculated based on the geometric center position of the test board socket and the geometric center position of the chip to be tested.

[0015] In addition, a preferred solution is that the operating method of the vision-based chip placement system provided by the present invention also includes: after the position of the chip to be tested is adjusted by the driving module based on the offset parameter, if the chip to be tested is still not placed in place, an error is reported through a preset error reporting module and the operation is terminated.

[0016] Compared with the prior art, the vision-based chip placement system and its operation method have the following beneficial effects: By setting up a position detection module linked to the driving module, it is possible to timely detect whether the chip to be tested is placed in place, and when the chip to be tested is not placed in place, it is possible to calculate the offset parameters of the chip to be tested. The driving module adjusts the position of the chip to be tested based on the offset parameters to put the chip to be tested in place, thereby significantly improving the accuracy of placing the chip into the test board socket and the work efficiency.

[0017] To achieve the above and related ends, one or more aspects of the present invention include specific features that will be described in detail below and particularly pointed out in the claims. The following description and the accompanying drawings set forth certain exemplary aspects of the present invention in detail. However, these aspects are merely indicative of the various ways in which the principles of the present invention may be employed. Furthermore, the present invention is intended to include all such aspects and their equivalents. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By referring to the following description and claims in conjunction with the accompanying drawings, and with a more complete understanding of the present invention, other objects and results of the present invention will become more apparent and readily understood. In the accompanying drawings: Figure 1 An enlarged view of a local structure of a vision-based chip placement system according to an embodiment of the present invention; Figure 2A flowchart of an operating method of a vision-based chip placement system according to an embodiment of the present invention; Figure 3 A comparison example diagram of a target detection model with a chip to be tested placed in place (corresponding to the left picture) and a comparison example of a target detection model with a chip to be tested not placed in place (corresponding to the right picture) according to an embodiment of the present invention; Figure 4 Diagrams marking the four corners of the test board socket for the target detection model provided by an embodiment of the present invention with the chip to be tested placed in place (corresponding to the left picture) and without the chip to be tested placed in place (corresponding to the right picture); Figure 5 Figure 1 shows the chip's four corners marked when the chip under test is properly placed (corresponding to the left picture) and when the chip under test is not properly placed (corresponding to the right picture) for the offset detection model provided according to an embodiment of the present invention; Figure 6 A diagram illustrating a method for determining an offset parameter according to an embodiment of the present invention; Reference numerals: robotic arm 1 , fixing belt 2 , cable alignment structure 3 , visual camera unit 4 , camera bracket 5 . DETAILED DESCRIPTION

[0019] In the following description, for illustrative purposes, numerous specific details are set forth to provide a comprehensive understanding of one or more embodiments. However, it will be apparent that the embodiments may be practiced without these specific details. In other instances, well-known structures and devices are shown in block diagram form to facilitate description of one or more embodiments.

[0020] The visual chip placement system provided by the present invention includes a driving module and a position detection module; wherein, The driving module is used to grab the chip to be tested and move it so as to initially place the chip to be tested on the test board socket; the position detection module is used to detect whether the chip to be tested is placed in place (whether the chip to be tested is completely placed in a preset area in the test board socket) when the chip to be tested is initially placed on the test board socket.

[0021] Furthermore, if the position detection module detects that the chip to be tested is not placed in place, the position detection module is also used to calculate the offset parameters of the chip to be tested, and the driving module is also used to adjust the position of the chip to be tested based on the offset parameters so that the chip to be tested is placed in place on the test board socket (the chip to be tested is completely placed in a preset area within the test board socket).

[0022] Specifically, Figure 1 The local structure magnification structure of the vision-based chip placement system according to an embodiment of the present invention is shown. Figure 1 It can be seen that in a specific embodiment of the present invention, the driving module may include a robot arm 1, which is used to grab the chip to be tested and shift the chip to be tested, so as to initially place the chip to be tested on the test board socket and adjust the position of the chip to be tested; and the position detection module includes a visual camera unit 4, which is arranged at the end of the robot arm, and the visual camera unit is used to capture visual data of the chip to be tested and the test board socket.

[0023] Further, if Figure 1 As shown, in order to realize the installation of the visual camera unit, a camera bracket 5 can be set at one end of the robot arm, the visual camera unit is set on the camera bracket 5, and the cable alignment structure 3 of the visual camera unit is fixed to the robot arm 1 through the fixing belt 2.

[0024] It should be noted that in the actual production process, the visual camera unit preferably uses an endoscope. An endoscope is a small periscope with both photo and video recording functions and has the advantage of being small. Due to the flexible mobility of the robotic arm, attaching the endoscope to the end of the robotic arm allows the endoscope to have a freer viewing angle than a fixed camera. The endoscope can be moved above the chip under test to observe the placement of the chip under test at an appropriate angle and distance. At the same time, the endoscope is connected to a control module (such as a computer) rather than directly to the robotic arm. Deploying the visual recognition model and algorithm on the control module side is more flexible and less costly.

[0025] Furthermore, it should be noted that, based on the camera bracket, the endoscope can rotate with the movement of the robotic arm. After the robotic arm completes the initial placement of the chip to be tested, it moves the endoscope to a preset photographing position. From this position, the endoscope can photograph the entire area of the chip to be tested and the test board socket from top to bottom at a vertical angle. The position detection module is used to control the endoscope to take photos of the chip to be tested and the test board socket as initial visual data of the chip to be tested and the test board socket.

[0026] In order to enable the position detection module to detect whether the chip to be tested is placed in place, the position detection module may further include a target detection model, which is used to determine whether the chip to be tested is placed in place based on the visual data of the chip to be tested and the visual data of the test board socket; specifically, when the visual data of the chip to be tested and the visual data of the test board socket meet a preset matching relationship, it is determined that the chip to be tested is placed in place; when the visual data of the chip to be tested and the visual data of the test board socket do not meet the preset matching relationship, it is determined that the chip to be tested is not placed in place.

[0027] In addition, in order to enable the position detection module to calculate the offset of the chip to be tested, the position detection module may also include an offset detection model; and when the chip to be tested is not placed in place, the offset detection model is used to calculate the offset parameters of the chip to be tested based on the visual data of the chip to be tested and the visual data of the test board socket.

[0028] It should be noted that, in order to realize the control of each module, the vision-based chip placement system provided by the present invention may also include a control module (such as a computer terminal), and the control module is electrically connected to the driving module and the position detection module respectively; in actual operation, the control module is used to send an instruction signal to the driving module to control the driving module to initially place the chip to be tested on the test board socket; the control module is also used to send an instruction signal to the position detection module to control the position detection module to take a picture of the chip to be tested and determine whether the chip to be tested is placed in place. When the chip to be tested is not placed in place, the position detection module calculates the offset parameters of the chip to be tested and sends the offset parameters to the control module; the control module is also used to send an instruction signal to the driving module based on the offset parameters to control the driving module to adjust the position of the chip to be tested so as to place the chip to be tested in place.

[0029] It should be noted that both the target detection model and the offset detection model can be built based on the YOLOv8 architecture (an existing artificial intelligence model architecture). The purpose of the target detection model is to identify whether the chip to be tested is placed correctly. The object classification model built using the YOLOv8 architecture can be used as the target detection model. This model can identify and classify objects from images and is easy to train, easy to deploy, and highly stable.

[0030] Furthermore, the target detection model adopted by the present invention is built based on the YOLOv8 (You Only Look Once version 8) architecture, which is a high-performance single-stage real-time target detection algorithm.

[0031] The structure of the target detection model mainly consists of three parts: Backbone: used to extract multi-level features of the input image. Common structures include CSPDarknet. Neck: Uses FPN (Feature Pyramid Network) or PAN (Path Aggregation Network) to fuse features of different scales to enhance the model's ability to detect objects of multiple sizes. Head: Outputs the prediction results, including the target center coordinates (x, y), width and height (w, h), confidence level, and category.

[0032] The output format is as follows: [x,y,w,h,confidence,class], where the coordinate values are normalized (ranging from 0 to 1).

[0033] Furthermore, the object detection model used in this invention uses a combined loss function for optimization during training, which mainly includes bounding box regression loss (such as IoU or CIoU), objectness loss (Objectness Loss), and classification loss (such as Binary Cross Entropy). The overall loss function is expressed as: Among them, each λ value is an adjustable weighting coefficient.

[0034] In addition, in order to train the target detection model, the present invention adopts a custom dataset. The original dataset contains a certain number (about 350) images, all of which have been manually labeled with target categories and bounding box positions (such as Figures 3 and 4 The dataset is divided into three parts: Training set (train): about 70%; Validation set (val): about 20%; Test set (test): about 10%.

[0035] All values are normalized coordinates and range between [0, 1].

[0036] In addition, in order to enhance the robustness of the target detection model and expand the number of training samples, the present invention applies multiple image enhancement methods to the training set, including brightness adjustment, exposure adjustment (gamma correction), and image rotation.

[0037] Brightness enhancement simulates different lighting environments by linearly scaling pixel values. The brightness variation range is set to ±15%, that is, the brightness factor α is randomly sampled from the interval , and its transformation formula is:

[0038] Exposure enhancement is achieved through Gamma correction, and the Gamma value λ is Randomly select from, the formula is:

[0039] Where λ<1 simulates overexposure (the image becomes brighter), λ>1 simulates underexposure (the image becomes darker), and I is the normalized pixel value.

[0040] Rotation enhancement is used to simulate the appearance changes of the target at different angles. The image is randomly rotated at an angle range of At the same time, the bounding box labels are updated synchronously with the affine transformation algorithm to ensure the accuracy of the enhanced labels.

[0041] Furthermore, augmentation operations can be combined, such as rotating an image and then adjusting its brightness and exposure to create diverse new samples. Through these augmentation processes, each original image can generate three to five augmented images, increasing the total number of images in the training set from approximately 245 to over 700, effectively alleviating the small sample size issue and improving model performance.

[0042] It should be noted that when training the target detection model, the training data is divided into two categories: correct and incorrect (see Figure 3 The standard for correctness or incorrectness is whether the chip can be tested after powering on). Mark the socket area frame as correct or incorrect (see Figure 4 ), the red box represents the frame range, and the red characters represent the recognized categories.

[0043] The purpose of the offset detection model provided by this invention is to calculate the offset parameters of the chip under test. These offset parameters mainly include the offset direction and the offset distance in the offset direction. The functions of the offset detection model provided by this invention mainly include two parts: identifying the actual position of the chip under test and the geometric center position of the test board socket.

[0044] Identifying the actual position of the chip under test primarily involves identifying the four corners of the chip under test. The offset detection model provided by the invention uses the YOLOv8 keypoint recognition model. Compared to the YOLOv8 object classification model, the YOLOv8 keypoint recognition model structure differs only in the head. The head outputs predictions, including the target center coordinates (x, y), width and height (w, h), object confidence, keypoint coordinates and confidence (kpt_n_x, kpt_n_y, kpt_n_score), and category.

[0045] The output format is as follows: [x,y,w,h,confidence, kpt_1_x,kpt_1_y,kpt_1_score … kpt_n_x,kpt_n_y,kpt_n_score class] The same training materials as the target detection model can be used for training. When marking, mark the chip frame as the chip and mark the four corners of the square chip as key points (see Figure 5 ), the red box represents the frame range, the red characters represent the classification of the recognized object, and the green dots represent the key points.

[0046] It should be noted that, usually, the test board socket has four fixing screws located at the four corners of the socket; and the geometric center of the square formed by the four fixing screws completely coincides with the geometric center of the square of the chip test area (preset area). Based on this, the bottoms of the four fixing screws can be marked (such as painted red with a marker). Through this setting, the four red dots (corresponding to the four fixing screws) can be identified from the green or metallic test board socket background by visual recognition, which can be used to subsequently determine the geometric center of the chip test area (preset area).

[0047] In addition, it should be noted that an image is composed of a set of pixels, and each pixel point p has an RGB value , the range is [0,255].

[0048] Assuming that the pixel set of the image is P, the red point set can be defined as:

[0049] in Represents the thresholds of red, green, and blue rgb values, divided into N connected subsets by connectivity:

[0050] Define the size (number of pixels) of each region as:

[0051] Set area threshold , then keep the area that meets the conditions:

[0052] By adjusting the RGB and area thresholds, we can finally filter out four areas, and define their centroids as the coordinates of the four red dots:

[0053] It should be noted that objects such as red jumpers and red short-circuit caps may be mistaken for red dots in the visual background. Since the scene is relatively fixed, in addition to setting the area threshold, users can use the following methods individually or in combination for screening: 1. Adjust RGB values: Under stable lighting conditions, even if the interfering element and the red dot are close in color, the two colors can still be distinguished by adjusting the RGB threshold due to the different reflective properties of different objects under lighting. The RGB threshold calculation formula is as follows:

[0054] 2. Set the aspect ratio threshold: Since all four reference points are circular, their shapes will be slightly distorted due to perspective, but the aspect ratio will remain close to 1:1. Setting the aspect ratio threshold can filter out interfering elements. This method is particularly effective for filtering objects such as jumpers.

[0055] like Figure 6 As shown, the geometric midpoint of the chip to be tested is obtained As shown by the yellow dots in the picture, the centers of the four screws As shown in the red dot position in the figure.

[0056] Will Figure 6 Put it into the rectangular coordinate system and convert from pixel coordinates to rectangular coordinates:

[0057] To simplify the description, the coordinates (x, y) involved in all the following formulas are in the rectangular coordinate system (with the lower left corner of the image as the origin and the y-axis pointing upward), and will not be noted otherwise.

[0058] set up In a straight line superior, In a straight line On, we get:

[0059] Right now

[0060] Output offset direction: like , then P is said to be located "above ”, like , then P is said to be located "below ”; Offset distance:

[0061] Output offset direction: like , then P is said to be on the "right side ” like , then P is said to be on the "left side ” Offset distance: in:

[0062] is the side length of the square formed by the red dots (can be measured with a vernier caliper).

[0063] With the above calculation method, the placement offset parameter of the chip under test can be obtained, and the position of the chip under test can be adjusted by the driving module based on the offset parameter to place the chip under test in place.

[0064] It should be noted that the above-mentioned offset parameter can reflect the relationship between the geometric center position of the chip to be tested and the geometric center position of the preset area of the test board socket. Under normal circumstances, the position adjustment of the chip to be tested can be completed by adjusting the position of the chip to be tested so that the geometric center of the chip to be tested coincides with the geometric center of the preset area of the test board socket. At this time, the chip to be tested can usually be located within the preset area of the test board socket. In addition, based on the structural features of the test board socket, such as the presence of transverse and longitudinal slope structures on the edge of the test board socket (not shown in the figure, which is a commonly used existing structure of the test board socket and has many records in existing literature, so it will not be repeated here), the chip to be tested can slide toward the correct position based on the slope structure; therefore, based on the presence of the transverse and longitudinal slope structures, the chip to be tested will usually only have transverse and longitudinal displacements during the movement process, and will not have movement in the rotational direction. Therefore, there is usually no need to perform angular rotation adjustment on the chip to be tested subsequently.

[0065] The vision-based chip placement system provided by this invention utilizes two YOLOv8 models and a visual camera unit to determine the correct placement of the chip under test and obtain information on correcting incorrect placement. This system utilizes only commonly available camera equipment, significantly reducing costs. Furthermore, the system can train both models using a limited dataset (approximately 300 images in a single environment), saving time and effort. Furthermore, the use of two different models leverages the high accuracy of the object recognition model and the information-richness of the key point recognition model, enabling precise identification and correction of the robot arm's placement of the chip under test.

[0066] On the other hand, to further illustrate the operation process of the vision-based chip placement system provided by the present invention, the present invention also provides an operation method of the aforementioned vision-based chip placement system, the method comprising: The chip to be tested is initially placed on the socket of the test board through the driving module; Detecting whether the chip to be tested is placed in place by a position detection module; If the chip to be tested is placed in place, the operation ends; If the chip under test is not in place, the position detection module calculates the offset parameter of the chip under test; and the driving module adjusts the position of the chip under test based on the offset parameter to put the chip under test in place, and then ends the operation.

[0067] Specifically, detecting whether the chip to be tested is placed in place by the position detection module includes: Detecting whether the visual data of the chip to be tested and the visual data of the test board socket meet a preset matching relationship; When the visual data of the chip to be tested and the visual data of the test board socket meet the preset matching relationship, it is determined that the chip to be tested is placed in place; when the visual data of the chip to be tested and the visual data of the test board socket do not meet the preset matching relationship, it is determined that the chip to be tested is not placed in place.

[0068] More specifically, calculating the offset parameter of the chip to be tested by the position detection module includes: Determining the geometric center position of the test board socket by visual data of the test board socket, and determining the geometric center position of the chip to be tested by visual data of the chip to be tested; The offset parameter is calculated based on the geometric center position of the test board socket and the geometric center position of the chip to be tested.

[0069] It should be noted that, in some special cases, when the chip to be tested still cannot be placed in place after position adjustment, the entire system needs to be stopped for maintenance; for this reason, the operating method of the vision-based chip placement system provided by the present invention also includes: after the position of the chip to be tested is adjusted by the driving module based on the offset parameter, if the chip to be tested is still not placed in place, an error is reported through the preset error reporting module, and the operation is terminated, and the entire system is subsequently repaired. As above Figures 1 to 6 The vision-based chip placement system and its operating method according to the present invention are described by way of example. However, those skilled in the art will appreciate that various modifications may be made to the vision-based chip placement system and its operating method without departing from the scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the contents of the appended claims.

Claims

1. A vision-based chip placement system, characterized in that: It includes a driving module and a position detection module; wherein, The driving module is used to initially place the chip to be tested on the test board socket; The position detection module includes a visual camera unit, which is used to capture visual data of the chip to be tested and the test board socket; The position detection module is used to detect whether the chip to be tested is placed in place based on the visual data of the chip to be tested and the visual data of the test board socket when the chip to be tested is initially placed on the test board socket; and If the chip to be tested is not placed in place, the position detection module is further used to calculate the offset parameter of the chip to be tested, and the driving module is further used to adjust the position of the chip to be tested based on the offset parameter to put the chip to be tested in place.

2. The vision-based chip placement system of claim 1, wherein: The driving module includes a mechanical arm, and the mechanical arm is used to grab the chip to be tested and shift the chip to be tested; and The visual camera unit is arranged on the robotic arm.

3. The vision-based chip placement system of claim 2, wherein: The position detection module further includes a target detection model, and the target detection model is used to determine whether the chip to be tested is placed in place based on the visual data of the chip to be tested and the visual data of the test board socket.

4. The vision-based chip placement system of claim 3, wherein: The position detection module further includes an offset detection model; and When the chip to be tested is not placed in place, the offset detection model is used to calculate the offset parameter of the chip to be tested based on the visual data of the chip to be tested and the visual data of the test board socket.

5. The vision-based chip placement system of claim 4, wherein: The target detection model and / or the offset detection model are built based on the YOLOv8 architecture.

6. The vision-based chip placement system of claim 5, wherein: It also includes a control module, which is electrically connected to the driving module and the position detection module respectively; and The control module is used to send instruction signals to the driving module and the position detection module.

7. A method for operating a vision-based chip placement system according to any one of claims 1 to 6, characterized in that: include: The chip to be tested is initially placed on the socket of the test board through the driving module; Detecting whether the chip to be tested is placed in place by a position detection module; If the chip to be tested is placed in place, the operation ends; If the chip under test is not in place, the position detection module calculates the offset parameter of the chip under test; and the driving module adjusts the position of the chip under test based on the offset parameter to put the chip under test in place, and then ends the operation.

8. The method for operating a vision-based chip placement system according to claim 7, wherein: Detecting whether the chip to be tested is placed in place by the position detection module includes: Detecting whether the visual data of the chip to be tested and the visual data of the test board socket meet a preset matching relationship; When the visual data of the chip to be tested and the visual data of the test board socket meet the preset matching relationship, it is determined that the chip to be tested is placed in place; when the visual data of the chip to be tested and the visual data of the test board socket do not meet the preset matching relationship, it is determined that the chip to be tested is not placed in place.

9. The method for operating a vision-based chip placement system according to claim 8, wherein: Calculating the offset parameter of the chip to be tested by the position detection module includes: Determining the geometric center position of the test board socket by visual data of the test board socket, and determining the geometric center position of the chip to be tested by visual data of the chip to be tested; The offset parameter is calculated based on the geometric center position of the test board socket and the geometric center position of the chip to be tested.

10. The method for operating a vision-based chip placement system according to claim 9, wherein: Also includes: After the position of the chip to be tested is adjusted by the driving module based on the offset parameter, if the chip to be tested is still not placed in place, an error is reported by a preset error reporting module and the operation is terminated.

Images