Bag placement detection system and method based on machine vision and storage medium

Abstract

The invention relates to the technical field of packaging machine bag placement, in particular to a bag placement detection system and method based on machine vision and a storage medium. Initial bag placement position images are collected, packaging bag types are matched through a target matching algorithm, and a packaging machine automatically sets parameters of new packaging bags; the multi-purpose automation degree of the packaging machine can be improved; by collecting an initial bag placing position image and a bag opening image and adopting an image processing method, detection of packaging bag size abnormity, packaging bag pattern abnormity, packaging bag opening abnormity, packaging bag empty and packaging feeding stopping is achieved, and then the packaging quality is improved. And meanwhile, image data acquisition is carried out by adopting an infrared and visible light dual-light camera, so that the empty packaging bag condition can be detected, an image segmentation mask can be established by virtue of infrared image data, and target detection, target classification and image processing carried out by using a neural network with a large calculated amount and a high computing power requirement are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of bag placing of packaging machines, and in particular to a bag placing detection system, method and storage medium based on machine vision. Background Art

[0002] Bag placement detection technology is widely used in packaging machinery in the food, medicine, daily chemical and other industries. Through accurate bag placement detection, it can ensure that each packaging bag is placed correctly without omission or overlap, thereby improving packaging quality and production efficiency. In the actual application of packaging machines, a packaging machine needs to be used for packaging of different products. When changing a new packaging bag, it is often necessary to reset the parameters and debug the packaging machine, which is time-consuming and labor-intensive and is not conducive to automated production. At the same time, after the packaging machine grabs the packaging bag, if the bag opening quality is not ideal, it will cause inaccurate material feeding or leakage, etc., which requires the operator to pay attention at all times, and the packaging quality is uncontrollable. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a bag placement detection system, method and storage medium based on machine vision. The system and method can automatically identify the current packaging bag and update the packaging parameters of the packaging machine, and at the same time have the function of detecting the bag opening effect of the packaging machine to prevent packaging quality problems caused by abnormal bag opening.

[0004] To achieve the above object, the present invention provides the following technical solution: a bag placement detection system based on machine vision, characterized in that the system comprises: The visual module is equipped with a camera to collect the image image1 of the initial bag placement position of the packaging machine; Data processing module: The data processing module receives the image data from the visual module. When the packaging machine is shut down and restarted, the data processing module first initializes the packaging bag data. The initialization process is as follows: (1) Receive the initial bag placement position image image1, use the image processing algorithm to segment the packaging bag target in image1, obtain the target image image0 and display it in the system display module; (2) The target matching algorithm is used to calculate the packaging bag type corresponding to image0. When the correct packaging bag type is not matched, the system display module issues a data not entered warning. When the packaging bag type is matched, the data processing module extracts the packaging bag specification data from the storage module; (3) Use image processing algorithms and combine camera lens parameters to calculate the current packaging bag specifications and dimensions. The actual size of the packaging bag is calculated as follows: , where L1 represents the calculated length of the packaging bag, W1 represents the calculated width of the packaging bag, R1 represents the calculated aspect ratio of the packaging bag, L2 represents the pixel length in the packaging bag image, and W2 represents the pixel width in the packaging bag image. represents the scale factor, where d represents the actual distance from the camera that takes image1 to the plane of the packaging bag, f represents the focal length of the camera, and m represents the pixel size of the camera CMOS; (4) By comparing the current type of packaging bag dimensions L0, W0, R0 stored in the storage module with the packaging bag specification dimensions L1, W1, R1 calculated through the image, when the judgment condition When the conditions are met, the bag is judged to be correctly matched, and the bag specification data obtained from the storage module is applied to the packaging machine, and the image data of image0 and the bag pixel sizes L2 and W2 are saved at the same time. When the judgment conditions are not met, the system display module issues a warning of abnormal bag size, where a, b, and c represent weights, and thresh0 represents the preset scale threshold; Display module: The display module displays the type and specification of the currently matched packaging bag on the one hand, and displays the operation process parameters on the other hand, and displays abnormal warning information at the same time, wherein the displayed operation process parameters include the number of currently packaged bags, the number of abnormal bags, the type of abnormality, etc.; Storage module, the storage module stores the user's packaging bag specification data, the camera calculation scale factor α, the initialized image0, L2, W2, and the data processing model and parameters.

[0005] Furthermore, when matching packaging bag types, if multiple types are matched, the packaging bag type selects the type with the highest matching score.

[0006] Furthermore, when the bag placement detection system is running, the visual module collects the initial bag placement position image image1_real in real time, and the data processing module receives image1_real for image segmentation, obtains image0_real and calculates the pixel size L of the packaging bag. 2-real , W 2-real ,when When , the system display module issues a warning of abnormal bag size, where β and γ represent weights, and thresh1 represents the preset pixel scale threshold.

[0007] Furthermore, when the bag placement detection system is running, the data processing module also compares the similarity between image0_real and image0 stored in the storage module. When the similarity comparison threshold is less than the specified threshold, the system display module issues a warning of abnormal pattern of the packaging bag, wherein the abnormal pattern of the packaging bag also includes abnormal pattern caused by damage to the packaging bag.

[0008] Furthermore, the visual module also collects the bag opening image image2, the data processing module receives image2 and uses an image processing algorithm to detect whether the bag opening is abnormal. When the detection result is that the bag opening is abnormal, the system display module issues a bag opening abnormality warning, and the bag placement detection system issues an instruction to discard the current bag. The bag opening abnormality includes that the bag is not opened or the bag opening gripper is detached.

[0009] Preferably, the camera carried by the vision module is an infrared and visible light dual-light camera, and the image data collected by the vision acquisition module includes a visible light image image_rgb and an infrared image image_r. When the packaging machine is initialized and running, the data processing module constructs a target image mask mask according to the infrared camera image image_r and performs image target segmentation on image_rgb based on the mask to obtain the segmented target image.

[0010] In particular, the visual module also collects an infrared image image3 of the bag body after filling. When the bag placement detection system is running, the data processing module also performs empty bag detection based on image3. When the detection result is an empty bag, the system display module issues an empty bag warning, and the bag placement detection system issues an instruction to discard the current bag.

[0011] Furthermore, when the data processing module detects that the packaging bag is empty N times in a row, the system display module issues a feeding stop warning, and the bag placement detection system issues a packaging machine pause instruction, where N is the preset maximum number of empty bags.

[0012] In particular, the present invention also provides a bag placement detection method based on machine vision, and the bag placement detection method includes packaging bag data initialization: packaging bag data initialization receives an initial bag placement position image image1, adopts an image segmentation algorithm to obtain target image data image0, adopts a target matching algorithm to obtain the packaging bag type, and when the correct packaging bag type is not matched, issues a data non-entry warning, and when the packaging bag type is matched, extracts the packaging bag specification data stored in the database, and at the same time, uses an image processing algorithm and combines the camera lens parameters to calculate the current packaging bag dimensions L1, W1, and R1, and compares them with the current type of packaging bag specification dimensions L0, W0, and R0 stored in the database. When If the matching is satisfied, the bag is judged to be correct, and the extracted bag specification data is applied to the packaging machine. At the same time, the image data of image0 and the bag pixel sizes L2 and W2 are saved in the database. If it is not met, a warning of abnormal bag size will be issued.

[0013] Furthermore, the bag placement detection method further includes a packaging bag size verification, including: receiving a real-time initial bag placement position image image1_real and segmenting it to obtain image0_real, calculating a packaging bag pixel size L 2-real , W 2-real ,when When the system displays the abnormal size of the packaging bag, the module will issue a warning.

[0014] Furthermore, the bag placement detection method also includes packaging bag pattern detection, including: performing a similarity comparison between image0_real and image0 stored in a database, and when a similarity comparison threshold is less than a specified threshold, issuing a packaging bag pattern abnormality warning, wherein the packaging bag pattern abnormality also includes a pattern abnormality caused by damage to the packaging bag.

[0015] Furthermore, the bag placement detection method also includes bag opening abnormality detection, including: receiving the bag opening image image2 in real time and using an image processing algorithm to detect whether the bag opening is abnormal; when the detection result is that the bag opening is abnormal, issuing a bag opening abnormality warning, wherein the bag opening abnormality includes that the bag is not opened or the bag opening gripper is detached.

[0016] Preferably, the bag detection method receives a visible light image image_rgb and an infrared image image_r. When performing bag detection, the bag detection method constructs a target image mask according to the infrared camera image image_r and performs image target segmentation on image_rgb based on the mask to obtain a segmented target image.

[0017] Furthermore, the bag placement detection method also includes empty bag detection, including: collecting an infrared image image3 of the bag body after filling and performing an empty bag detection on the bag, and when the detection result is an empty bag, the system issues an empty bag warning.

[0018] Furthermore, when the packaging bag empty bag detection detects that the packaging bag is empty for N consecutive times, a feeding stop warning is issued.

[0019] In particular, the present invention also provides a storage medium, which is a computer-readable storage medium for computer-readable storage, and stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the bag placement detection method based on machine vision.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention collects the initial bag placement position image, adopts the target matching algorithm to match the packaging bag type, realizes the packaging machine to automatically match the parameters of the new packaging bag and set the application, which can improve the automation level of the packaging machine for multiple uses; 2. The present invention collects the initial bag placement position image and the bag opening image, and realizes the detection of abnormal bag size, abnormal bag pattern, abnormal bag opening, empty bag, and stopped feeding of packaging bag through image processing method, which can further improve the packaging quality; 3. The present invention preferably uses an infrared and visible light dual-light camera for image data acquisition, which can not only detect the empty packaging bag, but also use the infrared image data to establish an image segmentation mask, thereby avoiding the use of a neural network with large computational complexity and high computing power requirements for target detection, target classification and image processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a functional module diagram of the bag placement detection system based on machine vision of the present invention; Figure 2 It is a logic diagram of the bag placement detection method based on machine vision of the present invention; Figure 3 The figure is a schematic diagram of an embodiment of the present invention for detecting abnormal bag opening of a packaging bag based on machine vision. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] like Figure 1 and Figure 2 As shown, the bag placement detection system and method based on machine vision provided by the present invention mainly solve the problem of automatic update of the operating parameters of the packaging machine when the packaging bags are replaced, and at the same time detect the problems of unqualified packaging machine size, abnormal packaging bag pattern, abnormal bag opening, empty bag and feeding stop caused by packaging bag squeezing, unqualified packaging bags, etc. during the operation of the packaging machine.

[0024] At the beginning of system installation, the visual module of the present invention needs to be installed. According to the actual size of the packaging machine and the field of view configuration of the visual camera, the visual camera can be one or more, as long as it can ensure that the visual camera can collect three types of images, namely: the initial bag placement position image image1, the bag mouth image image2 and the bag body infrared image image3 after filling, wherein the bag body image data collection position after filling can be specified during system debugging. After the visual module is installed, the distance d from the lens of the corresponding image1 collection to the plane where the initial position of the packaging bag is located, the camera focal length f, and the pixel size m will be determined and saved to the system database during system debugging, wherein the distance d may have errors, and the packaging bag itself has a height when it is placed, and these factors need to be considered when the threshold is finally judged. Furthermore, during system debugging, it is necessary to enter all the packaging bag specification parameters of the current user, such as the length, width and height of the packaging bag, the inflation volume, the sealing type, etc. Generally, the packaging bag type used by the packaging machine user is relatively fixed and the replacement frequency is low.

[0025] During operation, the packaging machine stops and starts, including just starting up and starting from a paused state. The system first collects image1 to initialize the packaging bag. On the one hand, it matches the packaging bag parameters so that the packaging machine automatically updates the corresponding parameters after replacing the packaging bag without manual adjustment. On the other hand, the packaging bag that is correctly matched is further image segmented to obtain image data containing only the packaging bag target after segmentation as a detection comparison sample for subsequent operation. Among them, the image segmentation algorithm can be a target detection based on a neural network, and then the original image is cropped according to the position parameters of the detection result to obtain the segmented data image0. As a preferred embodiment of the present invention, the present invention uses a dual-light camera for data acquisition, that is, infrared image data and visible light image data are collected simultaneously, a mask mask of the target packaging bag is established based on the infrared data, and then the target packaging bag is directly extracted in the visible light image according to the mask mask to obtain image0, thereby avoiding the problems of large amount of calculation and high computing power requirements caused by the use of neural networks. After image0 is obtained by image segmentation, a target matching algorithm can be used to further match the current packaging bag with the packaging bag type entered in the system database. The target matching algorithm here can adopt image classification or target detection based on neural networks, or target matching based on image similarity calculation. Matching the packaging bag type indicates that the packaging bag is stored in the database. When the target matches multiple types, that is, different sizes of the same pattern, the one with the highest matching degree is used by default in this embodiment. In subsequent applications, the present invention further determines whether the default match is correct based on size detection. When it is incorrect, it warns that the packaging bag size is abnormal and reminds manual reselection. At the same time, other matching items among the multiple types matched are provided as suggested selections to prevent the same pattern and different sizes of packaging from being matched.

[0026] The specific steps of the size-based detection are: obtaining the pixel sizes L2 and W2 of image0 through image processing and calculating the actual packaging bag size corresponding to image0 according to the camera imaging principle. The calculation formula is: , where ,when , it means that the previously matched packaging machine type is correct, and the corresponding packaging bag specification parameters are applied to the packaging machine system, and the image image0 and the packaging bag pixel size are saved. In actual application, because there is a measurement error on the one hand between the lens that collects image1 and the initial position plane of the packaging bag, and on the other hand, when the packaging bag is actually placed, the packaging bag has a thickness that will affect the accuracy of the d value. Therefore, in the judgment of whether the packaging size meets the conditions, c is given a greater weight, such as a=b=0.2, c=0.6, and thresh0 is preset according to the existing packaging bag size. Generally, the aspect ratio r of food packaging bags is in the range of 1~2. Considering the error e is 5%, then , when the balance is considered, thresh0=0.11 can be taken. In practical applications, when the aspect ratios of different packaging bags are quite different, different scale thresholds thresh0 corresponding to each type of packaging bag can be designed and stored in the corresponding packaging bag data. When the packaging bag is matched and the packaging bag specification parameters are extracted, the corresponding scale threshold thresh0 is extracted simultaneously; when the packaging bag is not matched, it means that the current packaging bag is not entered in the database or the matching is wrong. At this time, the system issues a warning and manual confirmation is performed according to the prompt information.

[0027] When the machine is restarted after shutdown, the system will initialize the packaging bags. After the initialization is completed, subsequent tests will continue to work on the basis of the initialization.

[0028] When working, the system of the present invention captures the initial bag placement position image image1_real in real time, and uses the initialized image segmentation method to obtain the current packaging bag image image0_real and the packaging bag pixel size L 2-real , W 2-real ,when in Represents the weight, usually , thresh1 represents the preset pixel scale threshold, such as 10, unit pixel, indicating that the size of the packaging bag may not be consistent with the actual size due to squeezing or the packaging bag itself is half missing, or the packaging bag has an extra piece. If the packaging is directly carried out at this time, it may affect the packaging quality. Therefore, the system issues a warning at this time, which is confirmed manually. In the application, the packaging gripper can also be directly driven to grab and discard the unqualified packaging bags to ensure the continuity of the packaging process.

[0029] Furthermore, the present invention also compares the similarity between the current packaging bag image image0_real and the image0 stored in the storage module. When the similarity value is less than the specified threshold, it indicates that there is a large difference between the texture of the current packaging bag and the texture initially stored. This difference may be a missing pattern, a disordered pattern, or even damage to the packaging bag itself. The system needs to issue a warning, wherein the similarity value comparison threshold is generally large, such as 0.5. In the application, the packaging gripper can also be directly driven to directly grab and discard unqualified packaging bags to ensure the continuity of the packaging process. In further applications, image0_real and image0 can be inter-frame differentiated to further indicate the location of the pattern difference, and the relevant results can be further uploaded.

[0030] Furthermore, in view of the possible problems that may exist in the process of the packaging machine grabbing the package and opening the bag, the present invention collects the bag opening image image2 and uses the image processing algorithm to detect whether the bag opening is abnormal. The algorithm can use a neural network to directly classify the abnormalities, or use an image contour tracking algorithm to track the connection area of ​​the bag. When the bag is not opened or one side of the gripper is not grasped after the bag is opened, the bag is separated from at least one side of the gripper. Therefore, the bag and the gripper are not connected during contour tracking. The final contour closed area is two large closed areas, while the closed area of ​​the normal bag opening bag is only one, such as Figure 3 As shown, normally there is only one closed area, so it is possible to complete the judgment of whether the bag opening is abnormal. When the detection result shows that the bag is not opened normally, the system display module issues a bag opening abnormality warning, and the bag placement detection system issues an instruction to discard the current bag.

[0031] Finally, when the filling of the packaging bag is completed, the infrared camera in this example also collects the infrared image image3 of the bag body after filling. Through the infrared picture, the image processing algorithm is used to detect the filler in image3 to determine whether there is filler in the current packaging bag. The image processing algorithm can use a neural network model for classification and judgment, or it can perform feature analysis on the infrared image to determine whether there is filler based on the color of the color block in the packaging bag in the infrared image. When the detection result is an empty bag, the system display module issues an empty bag warning, and the bag placement detection system issues an instruction to discard the current packaging bag; when the packaging bag is detected to be empty for N consecutive times (such as N=3), the system display module issues a feeding stop warning, indicating that the current packaging machine's unloading mechanism is empty and needs to be loaded first, and the bag placement detection system issues a packaging machine pause instruction.

[0032] It should be noted that when the packaging bag is a transparent bag, the recognition of the packaging may be affected. The corresponding processing algorithm should add image preprocessing to weaken the non-obvious features in the image. Typical methods include image corrosion, image sharpening, curve adjustment, etc. However, the above situation generally occurs less frequently. On the one hand, transparent bags are rarely used. On the other hand, the patterns on transparent bags have obvious features and can be detected.

[0033] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A bag placement detection system based on machine vision, characterized in that: The system includes: The visual module is equipped with a camera to collect the image image1 of the initial bag placement position of the packaging machine; Data processing module: The data processing module receives the image data from the visual module. When the packaging machine is restarted after stopping, the data processing module first initializes the packaging bag data, obtains the target image image0 containing only the packaging bag, and the pixel length L2 and pixel width W2 of the target packaging bag, and automatically sets the packaging machine parameters by judging the type of packaging bag; Display module: the display module displays the currently matched packaging bag type and specifications, and displays the operation process parameters and abnormal warning information; Storage module, the storage module stores the packaging bag specification data, the camera calculation scale factor α, the initialized image0, L2, W2, and the data processing model and parameters, among which the scale factor , d represents the actual distance from the camera that takes image1 to the plane of the packaging bag, f represents the focal length of the camera, and m represents the pixel size of the camera CMOS.

2. The bag placement detection system based on machine vision according to claim 1 is characterized in that: The steps for initializing the packaging bag data are as follows: (1) Receive the initial bag placement position image image1, use the image processing algorithm to segment the packaging bag target in image1, obtain the target image image0 and display it in the system display module; (2) The target matching algorithm is used to calculate the packaging bag type corresponding to image0. When the correct packaging bag type is not matched, the system display module issues a data not entered warning. When the packaging bag type is matched, the data processing module extracts the packaging bag specification data from the storage module; (3) Use image processing algorithm and camera lens parameters to calculate the current packaging bag specifications and dimensions. The calculation formula is: , L1 represents the calculated length of the packaging bag, W1 represents the calculated width of the packaging bag, R1 represents the calculated aspect ratio of the packaging bag, L2 represents the pixel length in the packaging bag image, and W2 represents the pixel width in the packaging bag image; (4) By comparing the current type of packaging bag specifications L0, W0, R0 stored in the storage module with L1, W1, R1 calculated through the image, when the judgment condition When the conditions are met, the bag is judged to be correctly matched. The system obtains the bag specification data from the storage module and applies it to the packaging machine. At the same time, the image data of image0 and the pixel sizes of the bag L2 and W2 are saved. When the judgment conditions are not met, the system display module issues a warning of abnormal bag size, where a, b, and c represent weights, and thresh0 represents the preset scale threshold.

3. The bag placement detection system based on machine vision according to claim 2 is characterized in that: When matching bag types, if multiple types are matched, the bag type with the highest matching score is selected.

4. The bag placement detection system based on machine vision according to claim 1 is characterized in that: When the bag placement detection system is running, the visual module collects the initial bag placement position image image1_real in real time, and the data processing module receives image1_real for image segmentation to obtain image0_real and calculate the pixel size L of the packaging bag. 2-real , W 2-real ,when When , the system display module issues a warning of abnormal bag size, where β and γ represent weights, and thresh1 represents the preset pixel scale threshold.

5. The bag placement detection system based on machine vision according to claim 4 is characterized in that: When the bag placement detection system is running, the data processing module also compares the similarity between image0_real and image0 stored in the storage module. When the similarity comparison threshold is less than the specified threshold, the system display module issues a warning of abnormal pattern of the packaging bag, wherein the abnormal pattern of the packaging bag also includes abnormal pattern caused by damage to the packaging bag.

6. The bag placement detection system based on machine vision according to claim 1 is characterized in that: The visual module also collects the bag opening image image2, the data processing module receives image2 and uses an image processing algorithm to detect whether the bag opening is abnormal. When the detection result is that the bag opening is abnormal, the system display module issues a bag opening abnormality warning, and the bag placement detection system issues an instruction to discard the current bag. The bag opening abnormality includes that the bag is not opened or the bag opening gripper is detached.

7. The bag placement detection system based on machine vision according to claim 1 is characterized in that: The camera carried by the vision module is an infrared and visible light dual-light camera. The image data collected by the vision acquisition module includes a visible light image image_rgb and an infrared image image_r. When the packaging machine is initialized and running, the data processing module constructs a target image mask mask according to the infrared camera image image_r and performs image target segmentation on image_rgb based on the mask to obtain the segmented target image image0.

8. The bag placement detection system based on machine vision according to claim 7 is characterized in that: The visual module also collects the infrared image image3 of the bag body after filling. When the bag placement detection system is running, the data processing module also performs empty bag detection based on image3. When the detection result is an empty bag, the system display module issues an empty bag warning, and the bag placement detection system issues an instruction to discard the current bag.

9. The machine vision-based bag placement detection system according to claim 8, characterized in that: When the data processing module detects that the packaging bag is empty N times in a row, the system display module issues a feeding stop warning, and the bag placement detection system issues a packaging machine pause instruction, where N is the preset maximum number of empty bags.

10. A bag placing detection method based on machine vision applied to the bag placing detection system based on machine vision according to any one of claims 1 to 9, characterized in that: The bag placement detection method includes packaging bag data initialization: the packaging bag data initialization receives the initial bag placement position image image1, adopts the image segmentation algorithm to obtain the target image data image0, adopts the target matching algorithm to obtain the packaging bag type, and when the correct packaging bag type is not matched, issues a data non-entry warning, and when the packaging bag type is matched, extracts the packaging bag specification data stored in the database, and at the same time, adopts the image processing algorithm and combines the camera lens parameters to calculate the current packaging bag size L1, W1, R1 and compares it with the current type of packaging bag specification size L0, W0, R0 stored in the database. If the matching is satisfied, the bag is judged to be correct, and the extracted bag specification data is applied to the packaging machine. At the same time, the image data of image0 and the bag pixel sizes L2 and W2 are saved in the database. If it is not met, a warning of abnormal bag size will be issued.

11. The bag placement detection method based on machine vision according to claim 10, characterized in that: The bag placement detection method also includes a packaging bag size verification, including: receiving a real-time initial bag placement position image image1_real and segmenting it to obtain image0_real, calculating the packaging bag pixel size L 2-real , W 2-real ,when When the system displays the abnormal size of the packaging bag, the module will issue a warning.

12. The bag placement detection method based on machine vision according to claim 11, characterized in that: The bag placement detection method also includes packaging bag pattern detection, including: performing a similarity comparison between image0_real and image0 stored in a database, and issuing a packaging bag pattern abnormality warning when a similarity comparison threshold is less than a specified threshold, wherein the packaging bag pattern abnormality also includes a pattern abnormality caused by a damaged packaging bag.

13. The bag placement detection method based on machine vision according to claim 10, characterized in that: The bag placement detection method also includes bag opening abnormality detection, including: receiving the bag opening image image2 in real time and using an image processing algorithm to detect whether the bag opening is abnormal, and issuing a bag opening abnormality warning when the detection result is that the bag opening is abnormal, wherein the bag opening abnormality includes that the bag is not opened or the bag opening gripper is detached.

14. The bag placement detection method based on machine vision according to claim 10, characterized in that: The bag detection method receives a visible light image image_rgb and an infrared image image_r. When performing bag detection, the bag detection method constructs a target image mask mask according to the infrared camera image image_r and performs image target segmentation on image_rgb based on the mask to obtain a segmented target image.

15. The bag placement detection method based on machine vision according to claim 14, characterized in that: The bag placement detection method also includes empty bag detection, including: collecting an infrared image image3 of the bag body after filling and performing an empty bag detection on the bag. When the detection result is an empty bag, the system issues an empty bag warning.

16. The bag placement detection method based on machine vision according to claim 15, characterized in that: When the packaging bag empty bag detection detects that the packaging bag is empty for N consecutive times, a feeding stop warning is issued.

17. A storage medium, the storage medium being a computer-readable storage medium, used for computer-readable storage, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the bag placement detection method based on machine vision as described in any one of claims 10-16.

Images