Workshop product package quality inspection system for baking industry

Abstract

The invention discloses a baking industry-oriented workshop product package quality inspection system. The system comprises an image acquisition module for acquiring a baked product package image; the preprocessing module is used for carrying out a series of preprocessing operations on the collected images and improving the image quality and the recognition accuracy; the defect detection module is used for automatically identifying and classifying packaging defects of the baked products based on a YOLOv7 target detection algorithm; the OCR character recognition module is responsible for recognizing printed characters on the package of the baked product; the result output and feedback module is used for integrating results of the defect detection module and the OCR character recognition module to generate a detailed quality inspection report; the technical problems that in the prior art, an automatic packaging quality inspection system for the baking industry is still scarce, and most systems do not fully consider the particularity and complexity of baking product packaging are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and in particular relates to a workshop product packaging quality inspection system for the baking industry. Background Art

[0002] In the baking industry, the quality of product packaging is directly related to product safety, shelf life and consumer experience. With the continuous expansion of the consumer market and the improvement of consumer requirements for product quality, the baking industry has an increasingly urgent need for workshop product packaging quality inspection.

[0003] Traditional bakery product packaging quality inspection methods mainly rely on manual visual inspection, but this method has many limitations. First, manual visual inspection is easily affected by subjective factors, such as worker fatigue and experience differences, which leads to inconsistency and instability in quality inspection results. Secondly, with the expansion of production scale and the improvement of the degree of automation of production lines, manual visual inspection can no longer meet the needs of fast and efficient quality inspection. In addition, for some subtle packaging defects, manual inspection may be difficult to detect, and there is a risk of missed detection.

[0004] However, there is still a lack of automated packaging quality inspection systems for the baking industry on the market. Traditional manual quality inspection methods have problems such as low efficiency, low accuracy, and susceptibility to subjective factors. Summary of the invention

[0005] The technical problem to be solved by the present invention is: to provide a workshop product packaging quality inspection system for the baking industry, so as to solve the technical problems that the automated packaging quality inspection systems for the baking industry in the prior art are still relatively scarce, and most systems do not fully consider the particularity and complexity of baking product packaging.

[0006] The technical solution of the present invention is: A workshop product packaging quality inspection system for the baking industry, the system comprising: Image acquisition module: obtains the image of bakery product packaging; Preprocessing module: performs a series of preprocessing operations on the collected images to improve image quality and recognition accuracy; Defect detection module: Automatically identify and classify defects in bakery product packaging based on the YOLOv7 target detection algorithm; OCR character recognition module: responsible for recognizing the printed text on the packaging of bakery products; Result output and feedback module: Integrate the results of the defect detection module and the OCR character recognition module to generate a detailed quality inspection report.

[0007] The image acquisition module is equipped with an industrial camera and a lighting system. The industrial camera captures the product packaging images on the production line in real time; the lighting system uses an evenly distributed light source to eliminate the effects of shadows and uneven lighting.

[0008] The preprocessing module performs a series of preprocessing methods on the collected images, including: first, removing noise and interference factors in the image through a denoising algorithm to reduce the impact on subsequent processing; then using an enhancement algorithm to improve the contrast and clarity of the image to make defects and key features more prominent; finally, using a binarization algorithm to convert the image into a black and white binary image to facilitate subsequent defect detection.

[0009] The YOLOv7 target detection algorithm uses the function of adaptively constraining the detection target size range and automatically adjusts the detection strategy according to the size of the target, thereby achieving the effect of accelerating detection and reducing false detections.

[0010] The defect detection process takes the preprocessed image, detection threshold and target intersection threshold as input to detect and locate various defects of product packaging, including damage, stains and misprints; the detection threshold is used to limit the detection of abnormal targets, and the target intersection threshold is used to limit the products to obtain a list of defect categories and positioning information [{cls1,pos1,score1},{cls2,pos2,score2},...,{clsn,posn,scoren}], where cls, pos and score represent the defect category information, positioning information and confidence of the product packaging respectively, and pos is in the form of [x1,y1,x2,y2], where x1, y1, x2 and y2 represent the four coordinates of the positioning information respectively.

[0011] The method for identifying printed text on the packaging of bakery products includes: firstly segmenting the image area containing text by using an image segmentation algorithm; then using an OCR engine to identify the segmented text area and convert it into text data; the OCR engine uses deep learning technology.

[0012] The quality inspection report is displayed to the user in a visual form, including the defect location, defect type, defect quantity, and information on the packaging.

[0013] The system also includes triggering warnings. When defects are detected in product packaging or key information is missing, the system immediately triggers the warning mechanism to remind the user through sound, light or pop-up window. The system also supports custom feedback settings. Users can set different feedback levels and prompt methods according to actual needs to ensure that quality inspection information is obtained at the first time. The system also supports data export function. Users can export quality inspection data as needed for more in-depth analysis and utilization.

[0014] Beneficial effects of the present invention: The system of the present invention includes multiple modules such as image acquisition, preprocessing, defect detection, OCR character recognition, and result output and feedback, which realizes comprehensive, efficient and automated detection of the packaging quality of baked products and significantly improves the product quality inspection effect in the baking industry.

[0015] First, the system uses high-resolution industrial cameras and optimized lighting systems to ensure that clear and stable images can be captured under various lighting conditions, providing a high-quality image data foundation for subsequent processing. Secondly, the preprocessing module performs preprocessing operations such as denoising, enhancement and binarization on the image, which effectively improves the image quality and recognition accuracy, and provides more favorable conditions for defect detection and OCR character recognition. The defect detection module adopts the advanced YOLOv7 target detection algorithm, which can maintain a real-time detection speed of more than 30FPS under high precision, greatly improving the detection efficiency. At the same time, its function of adaptively constraining the detection target size interval effectively reduces the false detection rate and improves the detection accuracy. The OCR character recognition module uses deep learning technology, which can accurately identify texts of various fonts, font sizes and typesetting methods, and convert them into editable text data, providing strong support for quality traceability and data analysis. Finally, the result output and feedback module displays the test results to the user in a visual form, allowing the user to intuitively understand the defects and key information of the product packaging, improving work efficiency and the convenience of quality inspection. At the same time, it supports data export function, which is convenient for users to conduct more in-depth analysis and utilization.

[0016] The present invention proposes a workshop product packaging quality inspection system for the baking industry. The system adopts advanced computer vision and deep learning technologies, combined with the characteristics of baking product packaging, to achieve automated and intelligent detection of packaging quality. The system not only improves the accuracy and efficiency of quality inspection and reduces labor costs, but also can adapt to different types and specifications of baking product packaging, with strong adaptability and flexibility. At the same time, the system is simple to operate and easy to maintain, providing a new solution for product packaging quality inspection in the baking industry.

[0017] The system of the present invention should be able to combine the characteristics of bakery product packaging, use advanced machine vision and deep learning technology, to achieve comprehensive, accurate and efficient detection of packaging quality, and provide strong support for quality improvement in the baking industry and improvement of consumer experience.

[0018] The present invention can realize the automatic and intelligent detection of product packaging, and improve the accuracy and efficiency of quality inspection. The image of product packaging is acquired by image acquisition equipment, and then the image is analyzed and processed by image processing algorithm, so as to realize the automatic recognition and classification of packaging defects. At the same time, the application of deep learning technology enables the system to continuously learn and optimize, and improve the accuracy and stability of recognition.

[0019] The invention solves the technical problems that the automated packaging quality inspection systems for the baking industry are still relatively scarce in the existing technology, and most of the systems do not fully consider the particularity and complexity of baking product packaging. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the system composition of the present invention. DETAILED DESCRIPTION

[0021] A workshop product packaging quality inspection system for the baking industry includes an image acquisition module, a preprocessing module, a defect detection module, an OCR character recognition module, and a result output and feedback module, which together realize a comprehensive, efficient, and automated inspection of the packaging quality of baking products. Figure 1 A framework design diagram of a workshop product packaging quality inspection system for the baking industry is given. In the figure, the image acquisition module is responsible for capturing the baking product packaging images on the production line in real time. The preprocessing module is responsible for preprocessing the acquired images to improve the image quality and recognition accuracy. The defect detection module uses the YOLOv7 target detection algorithm to detect and identify the preprocessed images to obtain the defect category information, positioning information and confidence of the product packaging. The OCR character recognition module is another important component of the system. It is responsible for identifying the text of various fonts, font sizes and typeset methods on the baking product packaging, and converting it into editable text data, providing strong support for quality traceability and data analysis. Finally, the result output and feedback module integrates the results of the feature extraction and defect detection module and the OCR character recognition module, and displays them to the user in a visual form, including defect location, defect type, defect quantity and key information on the packaging. The text similarity calculation system based on OCR invented by this patent has clear division of labor and works in coordination. The detailed description of each module is as follows: (1) Image acquisition module The image acquisition module is the first step for the system to acquire images of bakery product packaging. The module is equipped with a high-resolution industrial camera and an optimized lighting system to ensure that clear and stable images can be captured under various lighting conditions. The industrial camera has the characteristics of high speed, high precision and strong stability, and can capture product packaging images on the production line in real time. At the same time, the lighting system uses an evenly distributed light source to eliminate the influence of shadows and uneven lighting. In addition, the image acquisition device also has the ability to flexibly adjust to accommodate bakery product packaging of different sizes and shapes.

[0022] (2) Preprocessing module The preprocessing module performs a series of preprocessing operations on the collected images to improve image quality and recognition accuracy. First, the denoising algorithm is used to remove noise and interference factors in the image to reduce their impact on subsequent processing. Then, the enhancement algorithm is used to improve the contrast and clarity of the image to make defects and key features more prominent. Finally, the binarization algorithm is used to convert the image into a black and white binary image to facilitate subsequent defect detection. These preprocessing operations can significantly improve the image quality and recognition accuracy, and provide high-quality image data for subsequent processing.

[0023] (3) Defect detection module The defect detection module is the core part of the system. It realizes automatic identification and classification of defects in bakery product packaging based on the YOLOv7 target detection algorithm. The YoloV7 algorithm is currently the only detector that can maintain a real-time detection speed of more than 30FPS under high precision (optimal accuracy reaches 56.8% AP). This means that when processing large amounts of image or video data, YoloV7 can complete the detection task quickly and accurately, greatly improving work efficiency. The YoloV7 algorithm also has the function of adaptively constraining the detection target size range, and can automatically adjust the detection strategy according to the size of the target, thereby achieving the effect of accelerating detection and reducing false detection. It is more lightweight and efficient among the same type of detection algorithms, and is suitable for various complex scenarios and needs.

[0024] The defect detection process uses the preprocessed image, detection threshold and target intersection threshold as input to detect and locate various common defects of product packaging, such as damage, stains, misprints, etc. The detection threshold is used to limit the detection of abnormal targets, and the target intersection threshold is used to limit the products to obtain a list of defect categories and positioning information [{cls1,pos1,score1},{cls2,pos2,score2},...,{clsn,posn,scoren}], where cls, pos, and score represent the defect category information, positioning information, and confidence of the product packaging, respectively, and pos is in the form of [x1,y1,x2,y2], where x1, y1, x2, and y2 represent the four coordinates of the positioning information.

[0025] (4) OCR character recognition module The OCR character recognition module is responsible for identifying the printed text on the packaging of bakery products. This module uses advanced OCR technology, which can accurately identify text in various fonts, font sizes and layouts, and convert it into editable text data. First, the image area containing the text is segmented out through the image segmentation algorithm. Then, the segmented text area is recognized by the OCR engine and converted into text data. The OCR engine uses deep learning technology, has strong recognition capabilities and robustness, and can cope with various complex text recognition scenarios. Through the OCR character recognition module, the system can automatically read key information such as production date, batch number, product name, etc. on the packaging, providing strong support for subsequent quality traceability and data analysis.

[0026] (5) Result output and feedback module The result output and feedback module integrates the results of the defect detection module and the OCR character recognition module to generate a detailed quality inspection report. The report is presented to the user in a visual form, including the defect location, defect type, defect quantity, and key information on the packaging. Users can easily view and manage the test results through a friendly user interface without tedious operations. In addition, the system also provides a rich feedback mechanism. Once defects or missing key information in the product packaging are detected, the system will immediately trigger the early warning mechanism to remind the user through sound, light or pop-up windows. At the same time, the system also supports custom feedback settings. Users can set different feedback levels and prompt methods according to actual needs to ensure that important quality inspection information is obtained at the first time. In addition to real-time feedback, the system also supports data export function. Users can export quality inspection data as needed for more in-depth analysis and utilization. This module not only improves work efficiency, but also provides strong support for the quality management of baked products.

Claims

1. A workshop product packaging quality inspection system for the baking industry, characterized by: The system comprises: Image acquisition module: obtains the image of bakery product packaging; Preprocessing module: performs a series of preprocessing operations on the collected images to improve image quality and recognition accuracy; Defect detection module: Automatically identify and classify defects in bakery product packaging based on the YOLOv7 target detection algorithm; OCR character recognition module: responsible for recognizing the printed text on the packaging of bakery products; Result output and feedback module: Integrate the results of the defect detection module and the OCR character recognition module to generate a detailed quality inspection report.

2. According to claim 1, a workshop product packaging quality inspection system for the baking industry is characterized by: The image acquisition module is equipped with an industrial camera and a lighting system. The industrial camera captures the product packaging images on the production line in real time; the lighting system uses an evenly distributed light source to eliminate the effects of shadows and uneven lighting.

3. According to claim 1, a workshop product packaging quality inspection system for the baking industry is characterized by: The preprocessing module performs a series of preprocessing methods on the collected images, including: first, removing noise and interference factors in the image through a denoising algorithm to reduce the impact on subsequent processing; then using an enhancement algorithm to improve the contrast and clarity of the image to make defects and key features more prominent; finally, using a binarization algorithm to convert the image into a black and white binary image to facilitate subsequent defect detection.

4. According to claim 1, a workshop product packaging quality inspection system for the baking industry is characterized by: The YOLOv7 target detection algorithm uses the function of adaptively constraining the detection target size range and automatically adjusts the detection strategy according to the size of the target, thereby achieving the effect of accelerating detection and reducing false detections.

5. The workshop product packaging quality inspection system for the baking industry according to claim 1 is characterized by: The defect detection process takes the preprocessed image, detection threshold and target intersection threshold as input to detect and locate various defects of product packaging, including damage, stains and misprints; the detection threshold is used to limit the detection of abnormal targets, and the target intersection threshold is used to limit the products to obtain a list of defect categories and positioning information [{cls1,pos1,score1},{cls2,pos2,score2},...,{clsn,posn,scoren}], where cls, pos and score represent the defect category information, positioning information and confidence of the product packaging respectively, and pos is in the form of [x1,y1,x2,y2], where x1, y1, x2 and y2 represent the four coordinates of the positioning information respectively.

6. The workshop product packaging quality inspection system for the baking industry according to claim 1 is characterized by: The method for identifying printed text on the packaging of bakery products includes: firstly segmenting the image area containing text by using an image segmentation algorithm; then using an OCR engine to identify the segmented text area and convert it into text data; the OCR engine uses deep learning technology.

7. The workshop product packaging quality inspection system for the baking industry according to claim 1 is characterized by: The quality inspection report is displayed to the user in a visual form, including the defect location, defect type, defect quantity, and information on the packaging.

8. The workshop product packaging quality inspection system for the baking industry according to claim 1 is characterized by: The system also includes triggering an early warning. When a defect in the product packaging or missing key information is detected, the system immediately triggers an early warning mechanism to remind the user through sound, light or pop-up window; At the same time, the system also supports customized feedback settings. Users can set different feedback levels and prompt methods according to actual needs to ensure that quality inspection information is obtained as soon as possible. The system also supports data export function. Users can export quality inspection data as needed for more in-depth analysis and utilization.

Images