Package detection method and detection system

Abstract

The invention relates to the technical field of package detection, and provides a package detection method and system, and the method comprises the following steps: conveying a to-be-detected object along a curve track through a conveying device; when the to-be-detected object is conveyed to a detection station, shooting the to-be-detected object by using an image acquisition device to obtain at least two detection images in one-to-one correspondence with the image acquisition modules; analyzing the detection images, extracting reflective areas from the detection images, and fusing the detection images according to the reflective areas to obtain fused images corresponding to the detection stations; splicing all the fused images to form an overall image; and analyzing the overall image to generate a package detection result, and generating a conveying control instruction capable of controlling the conveying device according to the package detection result. The invention provides a package detection method and a detection system, which can be used for efficiently detecting a package and ensuring the accuracy of a detection result by eliminating a light reflecting area in a detection image.

Description

Technical Field

[0001] This invention relates to the field of packaging inspection technology, specifically a packaging inspection method and inspection system. Background Technology

[0002] Product packaging is a crucial part of the production process. Through reasonable structure and material selection, product packaging serves to protect the product, convey information, enhance brand value and user experience, while also meeting environmental protection requirements. To ensure the quality of packaging materials such as cardboard boxes and avoid defects such as misprints or damage that could affect the overall product quality, it is best to inspect the packaging materials during the production process.

[0003] With technological advancements, machine vision-based packaging inspection technology, with its high efficiency and precision, is gradually replacing traditional manual inspection techniques. Machine vision-based packaging inspection technology captures images of the packaging and then inspects these images to determine if there are defects such as misprints or damage. To ensure sufficient image clarity, supplementary lighting is usually used. However, after using supplementary lighting, many packaging images show bright reflective areas, affecting the accuracy of subsequent inspections.

[0004] To avoid the effects of reflective areas, most existing technologies focus on using machine learning to restore the reflective areas to their normal colors. However, machine learning is sensitive to image resolution. As image resolution increases, the complexity and training difficulty of machine learning models increase significantly, and the response speed in practical applications also slows down considerably. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a packaging inspection method and system that can efficiently inspect packaged goods and ensure the accuracy of inspection results by eliminating reflective areas in the inspection image.

[0006] To achieve the above objectives, the specific solution adopted by the present invention is as follows: a packaging inspection method, comprising the following steps: The object to be inspected is transported along a curved track using a conveyor device. The curved track has multiple inspection stations. An image acquisition device is set on the side of each inspection station. The image acquisition device includes at least two image acquisition modules arranged in parallel. Each image acquisition module includes an acquisition unit and a supplementary light. Different image acquisition devices have different shooting directions. When the object to be inspected is transported to the inspection station, the image acquisition device is used to capture images of the object to be inspected to obtain at least two inspection images that correspond one-to-one with the image acquisition module. The detection image is analyzed to extract the reflective area, and the detection image is fused according to the reflective area to obtain a fused image corresponding to the detection station; All merged images are stitched together to form the overall image; The overall image is analyzed to generate packaging inspection results, and based on the packaging inspection results, conveying control commands that can control the conveying device are generated.

[0007] As a further optimization of the above-mentioned packaging inspection method, the method for determining the inspection station on the curved trajectory includes: Multiple detection directions are determined based on the structural characteristics of the object to be detected; Select multiple direction change regions in the curve trajectory, and select an extended region near the direction change region; Set up detection stations in the direction change area and the extended area.

[0008] As a further optimization of the above-mentioned packaging inspection method: the image acquisition device includes a background plate, and when the object to be inspected is transported to the inspection station, the object to be inspected is located between the background plate and the image acquisition module.

[0009] As a further optimization of the above-mentioned packaging inspection method, the method for analyzing the inspection image to extract the reflective area includes: According to the preset sampling rules, multiple sample pixels are extracted from all the original pixels of the detected image and their RGB values ​​are determined. Highlight pixels are selected from sample pixels based on their RGB values; A spiral composed of multiple original pixels is constructed with the brightest pixel as the center. If the number of bright pixels in the spiral and the number of consecutive original pixels reaches a preset threshold, the radius of the region is obtained by calculating the distance between the last bright pixel in the spiral and the bright pixel at the center. A reflective area is generated based on the brightest pixel at the center and the radius of the area.

[0010] As a further optimization of the above-mentioned packaging detection method, a method for extracting multiple sample pixels from all original pixels of the detection image according to a preset sampling rule includes: Starting from the top left corner of the detected image, the detected image is divided into multiple matrix-distributed grids according to the preset grid size; Select the original pixel at the center point of the grid as the sample pixel.

[0011] As a further optimization of the above-mentioned packaging inspection method: the spiral is set as an Archimedean spiral, and when the spiral extends to the outside of the inspection image, the portion of the spiral located inside the inspection image is retained; When the reflective area extends outside the detection image, the portion of the reflective area located inside the detection image is retained.

[0012] As a further optimization of the above-mentioned packaging detection method: the method of fusing detection images according to the reflection area to obtain a fused image corresponding to the detection station includes: Setting all original pixels of the reflection area in the detection image as transparent; Superimposing all detection images corresponding to the image acquisition device, and performing mean value processing on all original pixels participating in superposition at each position to obtain a fused image corresponding to the detection station.

[0013] As a further optimization of the above-mentioned packaging detection method: the method of splicing all fused images to form an overall image includes: Determining the development plane of the detection image according to the structural features of the object to be detected, and dividing the development plane into multiple components; Establishing a mapping relationship between the fused image and the component according to the pose of the object to be detected when it is conveyed to the detection station; Splicing the fused image into an overall image according to the mapping relationship.

[0014] A packaging detection system for implementing the above-mentioned packaging detection method, the system includes a control device, a conveying device for conveying the object to be detected, and multiple image acquisition devices arranged on the side of the conveying device, and the control device is electrically connected with the conveying device and all image acquisition devices, the image acquisition device includes at least two parallel arranged image acquisition modules, the image acquisition module includes an acquisition device and a fill light, the shooting directions of different image acquisition devices are different.

[0015] As a further optimization of the above-mentioned packaging detection system: the image acquisition device includes a background plate, when the object to be detected is conveyed to the detection station, the object to be detected is located between the background plate and the image acquisition module, the background plate includes a color strip, the color strip has multiple color areas uniformly distributed along the length direction, and the colors of different color areas are different.

[0016] Beneficial effects: the application transports the detected object along the curved trajectory, so that the direction of the detected object can be changed when the detected object is transported to different positions, and then the different sides of the detected object can be photographed from the side of the detected object at the detection station, without actively rotating the detected object, which can reduce the complexity of the conveying device, reduce the control difficulty and the requirement for control accuracy of the conveying device, and is easier to implement; by photographing different sides of the detected object, defects at different positions on the packaging can be fully detected; compared with the conventional method of detecting the packaging first and then packaging the product, the application can detect the packaging of the packaged product, and can detect whether the packaging of the product with the packaging is damaged during transportation; after the detection image of the to-be-detected object is photographed, the reflection area is extracted from the detection image based on the RGB value of the original pixel in the detection image by constructing a spiral line, the number of original pixels to be processed is small, the complexity of the method can be reduced, and the efficiency of the method can be improved; the application fuses a plurality of detection images photographed by one image acquisition device on one side of the to-be-detected object based on the reflection area, which can quickly eliminate the reflection area in the detection image, and then ensure that the quality of the packaging can be accurately detected. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a flow chart of the detection method of the application; Figure 2 is a schematic diagram of the distribution mode of the image acquisition device; Figure 3 is a schematic diagram of the structure of the background plate; Figure 4 is a schematic diagram of the setting mode of the side limiting plate; Figure 5 is a schematic diagram of the setting mode of the rotating shaft and the extension plate.

[0018] BRIEF DESCRIPTION OF DRAWINGS: 1-conveying device, 2-to-be-detected object, 3-image acquisition module, 4-background plate, 5-housing cylinder, 6-center shaft, 7-end limiting plate, 8-connection shaft, 9-color band, 10-supporting plate, 11-side limiting plate, 12-closing plate, 13-rotating shaft, 14-extension plate. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0020] As Figure 1As shown, the present application firstly provides a packaging detection method, comprising S1 to S5.

[0021] S1, conveying the to-be-detected object along a curved track by using a conveying device, the curved track having a plurality of detection stations, image acquisition devices being arranged on the side of the detection stations, the image acquisition devices comprising at least two image acquisition modules arranged side by side, the image acquisition modules comprising an image collector and a light supplement lamp, the shooting directions of different image acquisition devices being different.

[0022] In the present application, the to-be-detected object is a packaging or a product with a packaging. Further, the present application is applicable to the detection of a carton and a product wrapped by the carton. The conveying device can adopt a conventional belt conveyor or a roller conveyor, etc., which are all mature prior art and will not be described here. Preferably, the belt conveyor is adopted, and the to-be-detected object will not rotate arbitrarily during the conveying process, the posture is stable as a whole, and the detection is easy. In the image acquisition module, the image collector is arranged as a camera with automatic zooming function, so that the to-be-detected object can be clearly shot. Such a camera is mature prior art and will not be described here. The light supplement lamp can adopt a white point light source, and the to-be-detected object can be illuminated by the light supplement lamp to ensure that the image collector can shoot a clear image.

[0023] More specifically, the method for determining the detection stations on the curved track comprises: determining a plurality of detection directions according to the structural features of the to-be-detected object; selecting a plurality of direction transformation regions in the curved track, and selecting an expansion region near the position of the direction transformation region; and arranging the detection stations in the direction transformation region and the expansion region.

[0024] By conveying the to-be-detected object along the curved track, the direction of the to-be-detected object can be changed when the to-be-detected object is conveyed to different positions, and then the different sides of the to-be-detected object can be shot from the side of the detection station, without actively rotating the to-be-detected object. This can reduce the complexity of the conveying device, the control difficulty and the requirement for control accuracy, and is easier to implement. On the other hand, by shooting different sides of the to-be-detected object, defects at different positions on the packaging can be fully detected. In addition, compared with the conventional method of detecting the packaging first and then packaging the product, the present application can detect the packaging of the product after the product has been packaged, and can detect whether the packaging of the product with the packaging is damaged during transportation.

[0025] S2, when the to-be-detected object is conveyed to the detection station, the image acquisition device is used to shoot the to-be-detected object to obtain at least two detection images corresponding to the image acquisition modules one by one.

[0026] S3, analyze the detection image to extract the highlight area therefrom, and fuse the detection image according to the highlight area to obtain a fused image corresponding to the detection station.

[0027] Specifically, the method of analyzing the detection image to extract the highlight area therefrom includes S31 to S35.

[0028] S31, extract a plurality of sample pixels from all original pixels of the detection image according to a preset sampling rule, and determine the RGB values of the sample pixels. Although the object to be detected can be illuminated by the light supplement lamp, a highlight area will be formed in the detection image, and the RGB values of the pixels in the highlight area will be abnormal, which will result in that the detection of whether the packaging has a misprint or other defects cannot be performed based on the detection image. Further, under the condition that the light supplement lamp is set as a white point light source, the highlight area is approximately circular, and the RGB values of the pixels in the highlight area will tend to the RGB values of white, i.e., tend to (255, 255, 255). Therefore, in S3, all original pixels are first analyzed based on the RGB values. When a high-definition collector is used, the number of original pixels of the detection image is large. In order to simplify the complexity of the method and improve the detection efficiency, a part of all original pixels is extracted as sample pixels in S31, and then only the RGB values of the sample pixels are determined.

[0029] Further, the method of extracting a plurality of sample pixels from all original pixels of the detection image according to a preset sampling rule includes S311 to S312.

[0030] S311, taking the upper left corner of the detection image as a starting point, divide the detection image into a plurality of grid cells according to a preset grid size.

[0031] S312, select the original pixel of the center point of the grid cell as a sample pixel.

[0032] Based on the extraction method, all sample pixels extracted are uniformly distributed in the detection image, and the position of the highlight area can be more accurately found based on the sample pixels in the subsequent process.

[0033] S32, screen out highlight pixels from the sample pixels according to the RGB values. Specifically, when the RGB values tend to (255, 255, 255), the sample pixels are determined as highlight pixels. The sample pixels can be screened by setting a color threshold, for example, setting the color preset value as (250, 250, 250). When the RGB values of the sample pixels exceed the color threshold, the sample pixels are determined as highlight pixels.

[0034] S33, constructing a spiral line composed of a plurality of original pixels with the highlight pixel as the center. More specifically, the spiral line is set as an Archimedes spiral, and when the spiral line extends to the outside of the detection image, the part of the spiral line located inside the detection image is reserved. As described above, the reflection area formed by the white point light source in the detection image is approximately circular, and if the highlight pixel belongs to the reflection area, there will be a large number of highlight pixels around the highlight pixel. Therefore, the highlight pixel is taken as the center, and the pixel values of the other original pixels around the highlight pixel are analyzed to determine whether the highlight pixel belongs to the reflection area. However, in a high-definition detection image, the number of original pixels around the highlight pixel is large, and if all of them are analyzed, there is still a problem of high complexity. Therefore, in S33, the spiral line is constructed with the highlight pixel as the center, and when the highlight pixel belongs to the reflection area, there will be many highlight pixels in the plurality of original pixels corresponding to the spiral line. The number of original pixels corresponding to the spiral line is significantly less than the number of original pixels around the center highlight pixel, so the complexity of the method can be reduced and the method is easier to implement.

[0035] It should be further noted that because all the original pixels in the detection image are matrix distributed, and the spiral line is a continuous line, in order to smoothly generate the spiral line, a plurality of reference points can be determined on the predetermined spiral line size, and when the reference point is located on one of the original pixels, the original pixel is included in the spiral line. If the reference point is located at the junction of different original pixels, one of the original pixels can be selected to be included in the spiral line.

[0036] S34, if the number of original pixels belonging to the highlight pixel and being continuous in front and back in the spiral line reaches a preset threshold, the distance between the last highlight pixel in the spiral line and the highlight pixel as the center is calculated to obtain the patch radius. If the center highlight pixel belongs to the reflection area and is close to the center of the reflection area, the part of the original pixels close to the center highlight pixel in the spiral line constructed based on the center highlight pixel will all be highlight pixels, and the part of the original pixels close to the end of the spiral line may not be highlight pixels. Therefore, based on the number of highlight pixels continuous in front and back, it can be determined whether the center highlight pixel belongs to the reflection area and whether it is close to the center of the reflection area. When it is determined that the center highlight pixel is close to the center of the reflection area, the distance between the original pixel at the end of the spiral line and the center highlight pixel can be used to calculate the patch radius, and the position and size of the reflection area can be estimated based on the position of the center highlight pixel and the patch radius.

[0037] S35, generate the highlight area according to the center highlight pixel and the radius of the segment. When at least two adjacent center highlight pixels belong to one highlight area and are close to the center of the highlight area, multiple highlight areas may be generated which overlap each other. In this case, the highlight area is generated according to the center highlight pixel with the largest number of highlight pixels in the corresponding spiral line. Further, when the highlight area extends to the outside of the detection image, the part of the highlight area located inside the detection image is reserved.

[0038] After the highlight area is determined, the method for fusing the detection images to obtain the fused image corresponding to the detection station includes S36-S37.

[0039] S36, set all original pixels of the highlight area in the detection image as transparent.

[0040] S37, superimpose all detection images corresponding to the image acquisition device, and perform mean value processing on all original pixels participating in superimposition at each position to obtain the fused image corresponding to the detection station. Because the image acquisition modules in the image acquisition device are arranged side by side, the highlight area in the detection image photographed by one of the image acquisition modules is normal in the detection image photographed by the other image acquisition module. After the original pixels of the highlight area are set as transparent, the different detection images are superimposed to obtain a clear fused image without the highlight area.

[0041] It should be further noted that if the packaging has a large number of white areas, these white areas may be determined as highlight areas. When superimposed, some positions may still be transparent after superimposition, which are the white areas in the packaging. In this case, the corresponding original pixels can be read from any detection image and added to the fused image. In this way, the white areas in the packaging can be avoided from being treated as highlight areas, thereby ensuring the integrity of the fused image.

[0042] S4, splice all the fused images to form the overall image.

[0043] Specifically, the method for splicing all the fused images to form the overall image includes S41-S43.

[0044] S41, determine the development plane of the detection image according to the structural features of the object to be detected, and divide the development plane into multiple components.

[0045] S42, establish a mapping relationship between the fused image and the components according to the pose of the object to be detected when it is conveyed to the detection station.

[0046] S43, splice the fused image into the overall image according to the mapping relationship.

[0047] In an embodiment of the present application, the packaging of the object to be detected is a square box, which has six sides, and correspondingly, the unfolded plane includes six components. For four of the sides, the image acquisition device arranged on the side of the curved track can be used for shooting, for the top side, the image acquisition device arranged above the curved track can be used for shooting, and for the bottom side, part of the position of the conveying device can be set to be transparent, and then the image acquisition device arranged below the curved track can be used for shooting, or a turnover mechanism can be added to the conveying device to detect the packaging after turning over. The turnover mechanism is a conventional structure in the conveying device, and will not be described here.

[0048] S5, analyze the overall image to generate a packaging detection result, and generate a conveying control instruction capable of controlling the conveying device according to the packaging detection result. When analyzing the overall image, the existing method for detecting whether the packaging has defects through machine learning technology can be used for analysis, such as the drug packaging box quality detection method and system based on a neural network model disclosed in Chinese patent document CN116703918A, the intelligent packaging detection method and system disclosed in Chinese patent document CN118195994A, or the intelligent production line packaging box detection system and method disclosed in Chinese patent document CN116935375A, etc., which will not be described here.

[0049] Considering that there are other background objects in the detection image in addition to the object to be detected, which can easily interfere with the detection process, in order to solve this problem, the image acquisition device includes a background plate, and when the object to be detected is conveyed to the detection station, the object to be detected is located between the background plate and the image acquisition module. By setting the background plate, the background part in the detection image can be quickly deleted, and only the part of the object to be detected is retained, thereby improving the efficiency of the method.

[0050] As shown in Figures 2 to 5 The present application further provides a packaging detection system for implementing the above packaging detection method, which includes a control device, a conveying device 1 for conveying the object to be detected, and a plurality of image acquisition devices arranged on the side of the conveying device, and the control device is electrically connected with the conveying device 1 and all the image acquisition devices. The image acquisition device includes at least two image acquisition modules 3 arranged side by side, the image acquisition module 3 includes an acquisition device and a fill light, and the shooting directions of different image acquisition devices are different.

[0051] Further, the image acquisition device comprises a background plate 4, when the object 2 to be detected is conveyed to the detection station, the object 2 to be detected is located between the background plate 4 and the image acquisition module 3, the background plate 4 comprises a color strip 9, the color strip 9 has a plurality of color areas uniformly distributed along the length direction, and the colors of different color areas are different. By setting the color strip 9, and setting a plurality of color areas with different colors on the color strip 9, a color area with obvious difference from the packaging object can be selected as the background based on the actual color of the packaging object, so that the background part in the detection image can be deleted, and the efficiency of the method is further improved.

[0052] The specific structure of the background plate 4 is that the background plate 4 comprises two containing barrels 5, the two containing barrels 5 are fixedly connected with a support plate 10, the two ends of the containing barrel 5 are each closed by a closing plate 12, a center shaft 6 is rotatably arranged at the center position of the containing barrel 5, the center shaft 6 is used for winding the color strip 9, two groups of end limiting plates 7 are fixedly connected on the center shaft 6, the end limiting plates 7 extend to the side of the center shaft 6, the two groups of end limiting plates 7 are arranged at intervals, and the color strip 9 is arranged between the two groups of end limiting plates 7, so that the color strip 9 is limited by the end limiting plates 7. An inlet and an outlet are formed on the barrel wall of the containing barrel 5, the color strip 9 can extend out of the inlet and outlet of one of the containing barrels 5 and enter from the inlet and outlet of the other containing barrel 5, by rotating the two center shafts 6, the position of the color strip 9 on the support plate 10 can be adjusted, so that the color areas with different colors are directed to the image acquisition device. In order to constrain the color strip 9, the support plate 10 is fixedly connected with at least one side limiting plate 11, and the side limiting plate 11 and the support plate 10 form a gap for the color strip 9 to pass through.

[0053] In order to facilitate the rotation of the center shaft 6, the center shaft 6 is coaxially fixedly connected with a connecting shaft 8, the connecting shaft 8 is fixedly connected with a rotating shaft 13 after penetrating through one of the closing plates 12, and the rotating shaft 13 is fixedly connected with two extension plates 14 extending in opposite directions.

[0054] The above description of disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A packaging inspection method, characterized in that, Includes the following steps: The object to be inspected is transported along a curved track using a conveyor device. The curved track has multiple inspection stations. An image acquisition device is set on the side of each inspection station. The image acquisition device includes at least two image acquisition modules arranged in parallel. Each image acquisition module includes an acquisition unit and a supplementary light. Different image acquisition devices have different shooting directions. When the object to be inspected is transported to the inspection station, the image acquisition device is used to capture images of the object to be inspected to obtain at least two inspection images that correspond one-to-one with the image acquisition module. The detection image is analyzed to extract the reflective area, and the detection image is fused according to the reflective area to obtain a fused image corresponding to the detection station; All merged images are stitched together to form the overall image; The overall image is analyzed to generate packaging inspection results, and based on the packaging inspection results, conveying control commands that can control the conveying device are generated.

2. The packaging inspection method as described in claim 1, characterized in that, Methods for determining the inspection station on a curved trajectory include: Multiple detection directions are determined based on the structural characteristics of the object to be detected; Select multiple direction change regions in the curve trajectory, and select an extended region near the direction change region; Set up detection stations in the direction change area and the extended area.

3. The packaging inspection method as described in claim 1, characterized in that, The image acquisition device includes a background plate. When the object to be inspected is transported to the inspection station, the object is located between the background plate and the image acquisition module.

4. The packaging inspection method as described in claim 1, characterized in that, The method for analyzing the detected image to extract the reflective area includes: According to the preset sampling rules, multiple sample pixels are extracted from all the original pixels of the detected image and their RGB values ​​are determined. Highlight pixels are selected from sample pixels based on their RGB values; A spiral composed of multiple original pixels is constructed with the brightest pixel as the center. If the number of bright pixels in the spiral and the number of consecutive original pixels reaches a preset threshold, the radius of the region is obtained by calculating the distance between the last bright pixel in the spiral and the bright pixel at the center. A reflective area is generated based on the brightest pixel at the center and the radius of the area.

5. The packaging inspection method as described in claim 4, characterized in that, Methods for extracting multiple sample pixels from all original pixels of a detection image according to preset sampling rules include: Starting from the top left corner of the detected image, the detected image is divided into multiple matrix-distributed grids according to the preset grid size; Select the original pixel at the center point of the grid as the sample pixel.

6. The packaging inspection method as described in claim 4, characterized in that, The spiral is configured as an Archimedean spiral, and when the spiral extends to the outside of the detection image, the portion of the spiral located inside the detection image is retained. When the reflective area extends outside the detection image, the portion of the reflective area located inside the detection image is retained.

7. The packaging inspection method as described in claim 1, characterized in that, Methods for fusing detection images based on reflective areas to obtain a fused image corresponding to the detection station include: Set all original pixels in the reflective areas of the detected image to transparent; All detection images corresponding to the image acquisition device are superimposed, and the average value of all original pixels participating in the superposition at each position is processed to obtain a fused image corresponding to the detection station.

8. The packaging inspection method as described in claim 1, characterized in that, Methods for stitching together all the merged images to form a single overall image include: The unfolding plane of the image to be detected is determined based on the structural features of the object to be detected, and the unfolding plane is divided into multiple components; Establish a mapping relationship between the fused image and its components based on the pose of the corresponding component when it is transported to the inspection station; The merged images are stitched together into a single overall image based on the mapping relationship.

9. A packaging inspection system, characterized in that, To implement a packaging inspection method as described in any one of claims 1-7, the system includes a control device, a conveying device (1) for conveying the object to be inspected, and a plurality of image acquisition devices disposed on the side of the conveying device. The control device is electrically connected to the conveying device (1) and all the image acquisition devices. The image acquisition device includes at least two image acquisition modules (3) arranged in parallel. The image acquisition module (3) includes an acquisition unit and a supplementary light. The shooting directions of different image acquisition devices are different.

10. A packaging inspection system as described in claim 9, characterized in that, The image acquisition device includes a background plate (4). When the object to be inspected (2) is transported to the inspection station, the object to be inspected (2) is located between the background plate (4) and the image acquisition module (3). The background plate (4) includes a color strip (9). The color strip (9) has multiple color areas that are evenly distributed along the length direction, and the colors of different color areas are different.

Images