Method and system for detecting a glass bottle

Abstract

A kind of detection method and system of glass bottle, including the following steps: glass bottle that passes on conveying belt is photographed, obtains with glass bottle detection image;Positioning is carried out to detection image, obtains the region where glass bottle is in detection image and detection frame is rotated to the region where glass bottle is;Image in detection frame is analyzed to determine whether glass bottle in detection frame exists defect, if there is defect, glass is marked as defective product, if there is no defect, the size of glass bottle is detected, whether it meets size threshold value is judged, if it does not meet size requirement, it is marked as defective product.By recognizing model to glass bottle and obtaining the region where glass bottle is, subsequently by algorithm, the identification frame is moved to the region where glass bottle is, to reduce background image, improve the success rate of glass bottle defect judgment.The present application carries out intelligent quality inspection to glass bottle by identification technology, reduces the cost of artificial detection, and also improves the efficiency of detection.

Description

Technical Field

[0001] This invention relates to the field of glass bottle testing technology, and in particular to a method and system for testing glass bottles. Background Technology

[0002] With social development, glass bottles are widely used in the beverage and pharmaceutical industries. During the manufacturing process, glass bottles are susceptible to defects due to unforeseen factors such as production processes. To ensure the quality of glass bottles, quality inspection is necessary before they leave the factory to remove substandard products. Currently, in the relevant production processes, glass bottles are inspected manually to remove substandard products from the production line. However, glass bottles are a product with high demand and large production volume, and manual inspection is inefficient and costly, which to some extent restricts production development. Summary of the Invention

[0003] To address the aforementioned shortcomings, the present invention aims to provide a method and system for inspecting glass bottles. This method utilizes identification technology to perform intelligent quality inspection of glass bottles, reducing the cost of manual inspection and improving inspection efficiency.

[0004] To achieve this objective, the present invention adopts the following technical solution: a method for detecting glass bottles, comprising the following steps: Step S1: Take pictures of the glass bottles passing by on the conveyor belt to obtain images with glass bottle detection; Step S2: Locate the detection image, obtain the area where the glass bottle is located in the detection image, and rotate the detection frame to the area where the glass bottle is located; Step S3: Analyze the image within the detection frame to determine whether the glass bottle within the detection frame is defective. If a defect is found, mark the glass as defective. If no defect is found, inspect the size of the glass bottle to determine whether it meets the size threshold. If the size requirement is not met, mark it as defective.

[0005] Preferably, the specific steps for determining whether the glass bottle within the detection frame has defects in step S3 are as follows: Step S31: Use the grayscale threshold segmentation algorithm to segment the image within the detection box to obtain at least two first sub-images. Obtain the first sub-image with a grayscale greater than the grayscale threshold, and judge the first sub-image through the first judgment parameter to obtain the first screening result. If the first screening result meets the requirements, proceed to step S32. Step S32: Use the dynamic threshold segmentation algorithm to segment the first sub-image after filtering to obtain several second sub-images. Obtain the second sub-image with the largest change in brightness, and judge the second sub-image through the second judgment parameter to obtain the second filtering result. If the second filtering result meets the requirements, proceed to step S33. Step S33: Use the linear feature search algorithm to search the first sub-image after filtering, obtain the suspicious area in the first sub-image, and judge the suspicious area through the third judgment parameter to obtain the third filtering result. If the third filtering result meets the requirements, the current glass bottle has no defects.

[0006] Preferably, before performing step S3, the detection image needs to be optimized. Image optimization includes one or more combinations of image filtering, enhancement algorithms, and denoising operations.

[0007] Preferably, the steps for inspecting the dimensions of the glass bottle are as follows: A model glass bottle of known size is photographed at different distances from the camera. The pixel size of the model glass bottle in the photos at different distances is obtained. The pixel distance ratio of the model glass bottle in the camera is obtained by using multiple sets of pixel sizes. The pixel distance ratio includes the pixel distance ratio in the horizontal direction and the pixel distance ratio in the vertical direction. Obtain the pixel dimensions of the first sub-image's length and width on the image. Using the pixel distance ratio and the pixel dimensions of the first sub-image's length and width on the image, obtain the true size of the first sub-image. Determine whether the true size meets the size threshold of the glass bottle.

[0008] A glass bottle detection system, using a glass bottle detection method, includes an imaging module, a positioning module, and a judgment module; The imaging module is used to capture images of glass bottles passing on the conveyor belt, thereby obtaining images with glass bottle detection. The positioning module is used to locate the detection image, obtain the area where the glass bottle is located in the detection image, and rotate the detection frame to the area where the glass bottle is located; The judgment module is used to analyze the image within the detection frame to determine whether the glass bottle within the detection frame is defective. If a defect is found, the glass is marked as a defective product. If no defect is found, the size of the glass bottle is detected to determine whether it meets the size threshold. If the size requirement is not met, it is marked as a defective product.

[0009] Preferably, the judgment module includes a defect judgment submodule; The defect judgment submodule includes a first judgment unit, a second judgment unit, and a third judgment unit; The first judgment unit is used to segment the image within the detection box using a grayscale threshold segmentation algorithm to obtain at least two first sub-images, obtain the first sub-image with a grayscale greater than the grayscale threshold, and judge the first sub-image through the first judgment parameter to obtain the first filtering result. If the first filtering result meets the requirements, the second judgment unit is called. The second judgment unit is used to segment the first sub-image after filtering using a dynamic threshold segmentation algorithm to obtain several second sub-images, to obtain the second sub-image with the largest change in brightness, and to judge the second sub-image by the second judgment parameter to obtain the second filtering result. If the second filtering result meets the requirements, the third judgment unit is called. The third judgment unit is used to search the first sub-image after screening using a linear feature search algorithm, obtain the suspicious area in the first sub-image, and judge the suspicious area through the third judgment parameter to obtain the third screening result. If the third screening result meets the requirements, the current glass bottle has no defects.

[0010] Preferably, it also includes an image preprocessing module, which is used to further optimize the detected image; Image optimization includes one or more combinations of image filtering, enhancement algorithms, and denoising operations.

[0011] Preferably, the judgment module includes a size judgment submodule, which is used to obtain the pixel size of the model glass bottle at different distances of the camera when the model glass bottle of known size is photographed, and obtain the pixel distance ratio of the model glass bottle in the camera through multiple sets of pixel sizes, wherein the pixel distance ratio includes the pixel distance ratio in the horizontal direction and the pixel distance ratio in the vertical direction. Obtain the pixel dimensions of the first sub-image's length and width on the image. Using the pixel distance ratio and the pixel dimensions of the first sub-image's length and width on the image, obtain the true size of the first sub-image. Determine whether the true size meets the size threshold of the glass bottle.

[0012] One of the above technical solutions has the following advantages or beneficial effects: By recognizing the glass bottle in the detection image and obtaining the area where the glass bottle is located, an algorithm moves the recognition box to the area where the glass bottle is located, thereby reducing background images and improving the success rate of defect detection for the glass bottle. Then, the glass bottle is judged for defects. If no defects are found after the glass bottle has been judged, its size is further judged. If the size also meets the requirements, the glass bottle is a good product; if any one of the requirements is not met, the glass bottle is judged as a defective product. This invention uses recognition technology to perform intelligent quality inspection of glass bottles, reducing the cost of manual inspection and improving inspection efficiency. Attached Figure Description

[0013] Figure 1 This is a flowchart of one embodiment of the method of the present invention.

[0014] Figure 2 This is a schematic diagram of the structure of one embodiment of the system of the present invention. Detailed Implementation

[0015] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0016] In the description of embodiments of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of embodiments of the present invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0017] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0018] like Figures 1-2 As shown, a method for testing glass bottles includes the following steps: Step S1: Take pictures of the glass bottles passing by on the conveyor belt to obtain images with glass bottle detection; Step S2: Locate the detection image, obtain the area where the glass bottle is located in the detection image, and rotate the detection frame to the area where the glass bottle is located; Step S3: Analyze the image within the detection frame to determine whether the glass bottle within the detection frame is defective. If a defect is found, mark the glass as defective. If no defect is found, inspect the size of the glass bottle to determine whether it meets the size threshold. If the size requirement is not met, mark it as defective.

[0019] In this invention, glass bottles are placed on a conveyor belt and transported along it. During transport, the bottles pass by photoelectric sensors, which control a camera to capture an image of the glass. A recognition model is then trained using existing models to identify the bottle and its location within the image. An algorithm moves the bounding box to this area to reduce background noise and improve defect detection accuracy. The bottle is then inspected for defects. If no defects are found, its dimensions are assessed. If the dimensions meet requirements, the bottle is considered good; otherwise, it is considered defective. This invention uses recognition technology for intelligent quality inspection of glass bottles, reducing the cost of manual inspection and improving efficiency.

[0020] Preferably, the specific steps for determining whether the glass bottle within the detection frame has defects in step S3 are as follows: Step S31: Use the grayscale threshold segmentation algorithm to segment the image within the detection box to obtain at least two first sub-images. Obtain the first sub-image with a grayscale greater than the grayscale threshold, and judge the first sub-image through the first judgment parameter to obtain the first screening result. If the first screening result meets the requirements, proceed to step S32. In defect identification, this invention first segments the detection image using a grayscale thresholding algorithm to obtain a first sub-image. This first sub-image contains both a background and the glass bottle itself. To clearly distinguish between the glass bottle and the background, the first sub-image is filtered based on a grayscale threshold. Different glass bottles require different grayscale thresholds. For example, when judging opaque glass bottles, because they are opaque, their grayscale value is lower than the surrounding environment in natural conditions. In this case, selecting the first sub-image with the highest grayscale value is sufficient to identify it as the glass bottle in the detection image. After filtering out the first sub-image, the glass bottle is identified, and further judgment is performed. The first judgment parameter can be the integrity of the glass bottle shape in the first sub-image. If the glass bottle shape in the first sub-image is incomplete, it indicates that the glass bottle may have a chip at the edge of the current shooting angle. In this case, the first screening result does not meet the requirements, and the glass bottle is considered defective.

[0021] Step S32: Use the dynamic threshold segmentation algorithm to segment the first sub-image after filtering to obtain several second sub-images. Obtain the second sub-image with the largest change in brightness, and judge the second sub-image through the second judgment parameter to obtain the second filtering result. If the second filtering result meets the requirements, proceed to step S33. When the glass bottle is intact, the next step of judgment is performed. Since step S31 judges the integrity of the glass bottle's edge shape from the current viewpoint, if a gap or irregular deformation appears in the middle of the glass bottle, step S31 cannot make a judgment. Therefore, in this invention, a dynamic threshold segmentation algorithm is used to segment the first sub-image after screening. The dynamic threshold in the dynamic threshold segmentation method can be brightness. When the glass bottle has defects or cracks, the light transmission under normal light is different from that of normal glass. Therefore, after segmenting multiple second sub-images, the multiple second sub-images are compared with each other to obtain the second sub-image with the largest change in brightness. It is then judged whether the brightness of this second sub-image meets the requirements of the second judgment parameter, which is the brightness parameter. If the brightness of this second sub-image does not meet the brightness parameter, the second screening result does not meet the requirements, and the glass bottle is a defective product.

[0022] Step S33: Use the linear feature search algorithm to search the first sub-image after filtering, obtain the suspicious area in the first sub-image, and judge the suspicious area through the third judgment parameter to obtain the third filtering result. If the third filtering result meets the requirements, the current glass bottle has no defects.

[0023] If a glass bottle has only a tiny crack, it has a very small impact on the light transmittance. Therefore, the method in step S32 cannot identify whether the glass bottle has a crack or scratch. In step S33, a linear feature search algorithm is used to identify and search the first sub-image after filtering, finding the line images present in the first sub-image and marking them as suspicious areas. Since the line images in suspicious areas are not necessarily cracks or scratches, this invention uses a third judgment parameter to judge the suspicious areas. This third judgment parameter is a model parameter for cracks or scratches. By comparing it with the line images in the suspicious areas, it can be determined whether it is a crack or a scratch. If the third filtering result meets the requirements, then the current glass bottle has no defects.

[0024] Preferably, before performing step S3, the detection image needs to be optimized. Image optimization includes one or more combinations of image filtering, enhancement algorithms, and denoising operations.

[0025] Preferably, the steps for inspecting the dimensions of the glass bottle are as follows: A model glass bottle of known size is photographed at different distances from the camera. The pixel size of the model glass bottle in the photos at different distances is obtained. The pixel distance ratio of the model glass bottle in the camera is obtained by using multiple sets of pixel sizes. The pixel distance ratio includes the pixel distance ratio in the horizontal direction and the pixel distance ratio in the vertical direction. Obtain the pixel dimensions of the first sub-image's length and width on the image. Using the pixel distance ratio and the pixel dimensions of the first sub-image's length and width on the image, obtain the true size of the first sub-image. Determine whether the true size meets the size threshold of the glass bottle.

[0026] The method for confirming the pixel distance of the model glass bottle is as follows: Using a model glass bottle of known size, take photos of the calibration object at camera distances of d=1, 2, 3, 4, and 5m. Use image processing tools to calculate the pixel distances of the width (x) and height (y) in the images. For example, if the actual width and height of the calibration object are 0.285m and 0.289m respectively, when the camera distance is d=1m, the pixel distances are x=200 pixels and y=300 pixels; when the camera distance is d=2m, the pixel distances are x=100 pixels and y=150 pixels. Since there is a linear relationship between pixel distance and actual distance, we can then calculate that when d=3m, x=66 pixels and y=100 pixels; when d=4m, x=50 pixels and y=75 pixels; and when d=5m, x=40 pixels and y=60 pixels. At a vertical distance of one meter, the actual distance represented by one pixel is: x = 0.285m ÷ 200, y = 0.289m ÷ 300. At a vertical distance of two meters, the actual distance represented by one pixel is: x = 0.285m ÷ 100 = 0.285 ÷ (200 / d), y = 0.289m ÷ 150 = 0.289 ÷ (300 / d). At a vertical distance of n meters, the pixel distance ratio of one pixel in the horizontal direction is: x = 0.285 ÷ (200 / n), and the pixel distance ratio of one pixel in the vertical direction is: y = 0.289 ÷ (300 / n). Since the pixel distance varies from camera to camera, it is necessary to measure the pixel distance using the above method before use.

[0027] During the detection process, since the glass bottle has already been extracted in step S31, the actual size of the glass bottle being photographed can be determined by the pixel size and pixel distance ratio in the horizontal and vertical directions of the first sub-image being filtered, thus achieving size detection.

[0028] A glass bottle detection system, using a glass bottle detection method, includes an imaging module, a positioning module, and a judgment module; The imaging module is used to capture images of glass bottles passing on the conveyor belt, thereby obtaining images with glass bottle detection. The positioning module is used to locate the detection image, obtain the area where the glass bottle is located in the detection image, and rotate the detection frame to the area where the glass bottle is located; The judgment module is used to analyze the image within the detection frame to determine whether the glass bottle within the detection frame is defective. If a defect is found, the glass is marked as a defective product. If no defect is found, the size of the glass bottle is detected to determine whether it meets the size threshold. If the size requirement is not met, it is marked as a defective product.

[0029] Preferably, the judgment module includes a defect judgment submodule; The defect judgment submodule includes a first judgment unit, a second judgment unit, and a third judgment unit; The first judgment unit is used to segment the image within the detection box using a grayscale threshold segmentation algorithm to obtain at least two first sub-images, obtain the first sub-image with a grayscale greater than the grayscale threshold, and judge the first sub-image through the first judgment parameter to obtain the first filtering result. If the first filtering result meets the requirements, the second judgment unit is called. The second judgment unit is used to segment the first sub-image after filtering using a dynamic threshold segmentation algorithm to obtain several second sub-images, to obtain the second sub-image with the largest change in brightness, and to judge the second sub-image by the second judgment parameter to obtain the second filtering result. If the second filtering result meets the requirements, the third judgment unit is called. The third judgment unit is used to search the first sub-image after screening using a linear feature search algorithm, obtain the suspicious area in the first sub-image, and judge the suspicious area through the third judgment parameter to obtain the third screening result. If the third screening result meets the requirements, the current glass bottle has no defects.

[0030] Preferably, it also includes an image preprocessing module, which is used to further optimize the detected image; Image optimization includes one or more combinations of image filtering, enhancement algorithms, and denoising operations.

[0031] Preferably, the judgment module includes a size judgment submodule, which is used to obtain the pixel size of the model glass bottle at different distances of the camera when the model glass bottle of known size is photographed, and obtain the pixel distance ratio of the model glass bottle in the camera through multiple sets of pixel sizes, wherein the pixel distance ratio includes the pixel distance ratio in the horizontal direction and the pixel distance ratio in the vertical direction. Obtain the pixel dimensions of the first sub-image's length and width on the image. Using the pixel distance ratio and the pixel dimensions of the first sub-image's length and width on the image, obtain the true size of the first sub-image. Determine whether the true size meets the size threshold of the glass bottle.

[0032] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0033] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for testing glass bottles, characterized in that, Includes the following steps: Step S1: Take pictures of the glass bottles passing by on the conveyor belt to obtain images with glass bottle detection; Step S2: Locate the detection image, obtain the area where the glass bottle is located in the detection image, and rotate the detection frame to the area where the glass bottle is located; Step S3: Analyze the image within the detection frame to determine whether the glass bottle within the detection frame is defective. If a defect is found, mark the glass as defective. If no defect is found, inspect the size of the glass bottle to determine whether it meets the size threshold. If it does not meet the size requirements, mark it as defective. The specific steps for determining whether the glass bottle in the detection frame has defects in step S3 are as follows: Step S31: Use the grayscale threshold segmentation algorithm to segment the image within the detection box to obtain at least two first sub-images. Obtain the first sub-image with a grayscale greater than the grayscale threshold, and judge the first sub-image through the first judgment parameter to obtain the first screening result. If the first screening result meets the requirements, proceed to step S32. Step S32: Use the dynamic threshold segmentation algorithm to segment the first sub-image after filtering to obtain several second sub-images. Obtain the second sub-image with the largest change in brightness, and judge the second sub-image through the second judgment parameter to obtain the second filtering result. If the second filtering result meets the requirements, proceed to step S33. Step S33: Use the linear feature search algorithm to search the first sub-image after filtering, obtain the suspicious area in the first sub-image, and judge the suspicious area through the third judgment parameter to obtain the third filtering result. If the third filtering result meets the requirements, the current glass bottle has no defects.

2. The method for detecting glass bottles according to claim 1, characterized in that, Before proceeding to step S3, image optimization is required for the detected image; Image optimization includes one or more combinations of image filtering, enhancement algorithms, and denoising operations.

3. The method for detecting glass bottles according to claim 1, characterized in that, The steps for inspecting the dimensions of glass bottles are as follows: A model glass bottle of known size is photographed at different distances from the camera. The pixel size of the model glass bottle in the photos at different distances is obtained. The pixel distance ratio of the model glass bottle in the camera is obtained by using multiple sets of pixel sizes. The pixel distance ratio includes the pixel distance ratio in the horizontal direction and the pixel distance ratio in the vertical direction. Obtain the pixel dimensions of the first sub-image's length and width on the image. Using the pixel distance ratio and the pixel dimensions of the first sub-image's length and width on the image, obtain the true size of the first sub-image. Determine whether the true size meets the size threshold of the glass bottle.

4. A glass bottle detection system, using the glass bottle detection method according to any one of claims 1 to 3, characterized in that, It includes a shooting module, a positioning module, and a judgment module; The imaging module is used to capture images of glass bottles passing on the conveyor belt, thereby obtaining images with glass bottle detection. The positioning module is used to locate the detection image, obtain the area where the glass bottle is located in the detection image, and rotate the detection frame to the area where the glass bottle is located; The judgment module is used to analyze the image within the detection frame to determine whether the glass bottle within the detection frame is defective. If a defect is found, the glass is marked as a defective product. If no defect is found, the size of the glass bottle is detected to determine whether it meets the size threshold. If the size requirement is not met, it is marked as a defective product.

5. The glass bottle detection system according to claim 4, characterized in that, The judgment module includes a defect judgment submodule; The defect judgment submodule includes a first judgment unit, a second judgment unit, and a third judgment unit; The first judgment unit is used to segment the image within the detection box using a grayscale threshold segmentation algorithm to obtain at least two first sub-images, obtain the first sub-image with a grayscale greater than the grayscale threshold, and judge the first sub-image through the first judgment parameter to obtain the first filtering result. If the first filtering result meets the requirements, the second judgment unit is called. The second judgment unit is used to segment the first sub-image after filtering using a dynamic threshold segmentation algorithm to obtain several second sub-images, to obtain the second sub-image with the largest change in brightness, and to judge the second sub-image by the second judgment parameter to obtain the second filtering result. If the second filtering result meets the requirements, the third judgment unit is called. The third judgment unit is used to search the first sub-image after screening using a linear feature search algorithm, obtain the suspicious area in the first sub-image, and judge the suspicious area through the third judgment parameter to obtain the third screening result. If the third screening result meets the requirements, the current glass bottle has no defects.

6. The glass bottle detection system according to claim 5, characterized in that, It also includes an image preprocessing module, which is used to further optimize the detected image; Image optimization includes one or more combinations of image filtering, enhancement algorithms, and denoising operations.

7. The glass bottle detection system according to claim 6, characterized in that, The judgment module includes a size judgment submodule, which is used to obtain the pixel size of the model glass bottle at different distances of the camera when the model glass bottle is photographed with a known size. The pixel size of the model glass bottle at different distances is obtained, and the pixel distance ratio of the model glass bottle in the camera is obtained by multiple sets of pixel sizes. The pixel distance ratio includes the pixel distance ratio in the horizontal direction and the pixel distance ratio in the vertical direction. Obtain the pixel dimensions of the first sub-image's length and width on the image. Using the pixel distance ratio and the pixel dimensions of the first sub-image's length and width on the image, obtain the true size of the first sub-image. Determine whether the true size meets the size threshold of the glass bottle.

Images