Image self-correction intelligent scanning platform and sorting optimization method

Abstract

The invention provides an image self-correction intelligent scanning platform and a sorting optimization method, and relates to the technical field of image processing, the platform comprises a multi-angle image acquisition module, a bar code restoration and identification module, an image evaluation module and an image classification module. The scanning device is used for obtaining image information of an article to be recognized and comprises a bar code image and an article sorting module. An AI repair algorithm with a fuzzy or damaged bar code is built in the bar code repair and recognition module, the fuzzy or damaged bar code image obtained through scanning can be repaired, the problems of adhesive tape coverage and recognition of stained and damaged bar codes are solved, and the code scanning recognition rate is increased. In the scanning process, goods types can be automatically marked, manual labeling classification is replaced, the labor cost of a sorting center is reduced, efficient and accurate logistics sorting optimization is achieved, and the problem of the sorting efficiency bottleneck caused by bar code failure in logistics storage is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to an intelligent scanning platform for image self-correction and a sorting optimization method. Background Art

[0002] In the operation of logistics warehouses, scanning, identification, classification and sorting of goods are key links. However, traditional scanning equipment and sorting methods face many challenges. On the one hand, barcodes on goods are easily affected by factors such as tape coverage and contamination, resulting in a reduced scanning and recognition rate, which affects the efficiency of goods circulation. On the other hand, manual labeling and sorting are not only inefficient but also have high labor costs. With the rapid development of the logistics industry, the volume of goods handled continues to increase, and the requirements for sorting efficiency and accuracy are becoming increasingly higher. Existing technical means are difficult to meet this demand: for example, multispectral scanning is used to enhance barcode recognition, but it cannot reconstruct the structure of damaged barcodes; barcode repair based on template matching has poor generalization ability and cannot adapt to complex contamination scenarios; another method is to replace barcodes with RFID, but the labeling cost increases by 200%, making it difficult to apply on a large scale.

[0003] Therefore, there is an urgent need for an intelligent scanning and sorting optimization solution that can improve barcode recognition rate and reduce labor costs. Summary of the Invention

[0004] To solve the above problems, the present invention provides an intelligent scanning platform with self-correction of images and a sorting optimization method. Through the AI ​​repair algorithm for blurred and damaged barcodes built into the barcode repair module, the blurred or damaged barcode images obtained by scanning can be repaired, solving the problem of identifying barcodes covered with tape or stained, and improving the scanning recognition rate; the image classification module has a built-in AI classification model for object images, which can automatically mark the type of goods during the scanning process, replacing manual labeling and classification, and reducing the labor cost of the sorting center.

[0005] To achieve the above object, the technical solution adopted by the present invention is: In a first aspect of the present invention, an intelligent scanning platform for image self-correction is provided, comprising: A multi-angle image acquisition module is configured above and to the side of the sorting line transport device, and is used to dynamically acquire image information of the items to be identified, including barcode images and images of the item's appearance; A barcode repair and recognition module is connected to the multi-angle image acquisition module. The barcode repair module has a built-in barcode AI repair algorithm for repairing the collected blurred or damaged barcode images and identifying the items based on the repaired barcodes. An image classification module, connected to the multi-angle image acquisition module, having a built-in AI classification model for item images, for automatically marking the type of goods based on the barcode image and the item appearance image during the scanning process; An image evaluation module is connected to the barcode repair and recognition module and the image classification module, and is used to integrate the barcode recognition results and the AI ​​classification label to identify the type of item; The item sorting module is connected to the image evaluation module and is used to generate sorting path instructions according to the item type and link the sorting robot arm to execute them.

[0006] Preferably, the multi-angle image acquisition module includes an acquisition device arranged above and on the side of the sorting line transport device, and the acquisition device is a high-definition industrial camera. The image information collected by the high-definition industrial camera is then subjected to basic preprocessing, and the preprocessing steps include Gaussian smoothing noise reduction and histogram equalization to enhance contrast.

[0007] Preferably, the AI ​​repair algorithm includes: Recognition unit, used to identify the blurred or damaged area in the barcode image and determine its position and range; The image repair model is trained using a large number of blurred and damaged barcode images and their corresponding standard clear barcode images. The blurred or damaged barcode images are then repaired through the trained deep learning model to generate repaired clear barcode images, and objects are identified based on the repaired clear barcode images.

[0008] More preferably, the restoration result is evaluated by introducing an image restoration accuracy index, and the calculation formula of the image restoration accuracy index is as follows: , in, represents the image restoration accuracy index, represents the coefficient of variation, represents the ground-corrected image with the coordinates (i, j) on the image, Represents the output image at position (i, j) on the image, Indicates Centered along its main migration direction The extended linear neighborhood window, the greater the degree of deviation, the greater the fluctuation of the shadow of the point in its direction neighborhood, which does not conform to the principle of diffusion path stability. N represents the number of samples, the image restoration accuracy index The closer the value is to 0, the higher the repair accuracy is.

[0009] Preferably, the object image AI classification model includes: Feature extraction unit, used to extract key features from the object appearance image, including shape, color and texture; Classification training unit, including deep learning classification models, is trained using a large number of labeled item images to accurately identify different types of goods; The classification execution unit is used to classify and identify the newly scanned appearance image of the item based on the trained deep learning classification model, automatically mark the type of goods, generate classification result information, and store it in association with the corresponding barcode information.

[0010] Preferably, the image evaluation module recognizes the type of the item by the following steps: When the barcode recognition confidence is greater than 90%, the barcode information is given priority; When the barcode recognition confidence level is less than or equal to 90%, a weighted decision is made based on the item classification results and barcode information. The weighted decision is calculated by the following formula: , in, represents a weighted decision, is the weighting coefficient, is the barcode recognition result of the item, is the image recognition result of the object.

[0011] Preferably, the item sorting module comprises a sorting robot arm, which sorts items to corresponding target areas according to item types.

[0012] In a second aspect of the present invention, a sorting optimization method is further provided, comprising the following steps: S1: High-definition industrial cameras installed above and on the sides of the sorting line transport device dynamically capture multi-view images of items in transport; S2: Evaluate the integrity of the barcode area. If it is blurred or damaged, the barcode will be repaired and recognized using the barcode AI repair algorithm. S3: Parallel execution of dual-channel analysis of barcode recognition and object image classification, where the object image classification is achieved using an image AI classification model; S4: Combine barcode recognition results and AI classification labels to identify the item type; S5: Map the recognition results to the sorting instruction set to drive the sorting robot arm to automatically sort the items.

[0013] The beneficial effects of the present invention are: 1. Improve the barcode recognition rate: The built-in AI repair algorithm for blurred and damaged barcodes can effectively solve the recognition problem of tape-covered and stained barcodes in logistics warehouses, significantly improve the barcode recognition rate, ensure the accurate acquisition of cargo information, reduce cargo processing delays caused by barcode problems, and improve the overall operational efficiency of logistics warehouses.

[0014] 2. Reduce labor costs: Utilizing the built-in AI classification model for item images, the system automatically labels the type of goods while scanning them, replacing traditional manual labeling and classification work. This greatly reduces the manpower input and labor costs of the sorting center, while improving the accuracy and consistency of classification and avoiding errors that may occur in manual classification.

[0015] 3. Optimize the sorting process: The repaired barcode information and automatically marked cargo type information are promptly sent to the sorting control system. The sorting control system can accurately plan sorting routes and assign tasks based on this information, thereby achieving efficient and accurate logistics sorting optimization, improving sorting efficiency, reducing the residence time of cargo in the warehouse, speeding up the turnover of cargo, and improving customer satisfaction.

[0016] 4. Improvement of intelligence and automation: This invention deeply integrates AI technology into logistics scanning and sorting links, realizes image self-correction and automatic classification functions, promotes the intelligent and automated development of logistics warehouses, provides strong support for the digital transformation of the logistics industry, and helps to improve the competitiveness and market adaptability of logistics companies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a structural block diagram of an intelligent scanning platform for image self-correction according to the present invention.

[0018] Figure 2 This is a flow chart of a sorting optimization method of the present invention. DETAILED DESCRIPTION

[0019] See also Figure 1 As shown, in a first aspect of the present invention, an intelligent scanning platform for image self-correction is provided, comprising: A multi-angle image acquisition module is configured above and to the side of the sorting line transport device, and is used to dynamically acquire image information of the items to be identified, including barcode images and images of the item's appearance; A barcode repair and recognition module is connected to the multi-angle image acquisition module. The barcode repair module has a built-in barcode AI repair algorithm for repairing the collected blurred or damaged barcode images and identifying the items based on the repaired barcodes. An image classification module, connected to the multi-angle image acquisition module, having a built-in AI classification model for item images, for automatically marking the type of goods based on the barcode image and the item appearance image during the scanning process; An image evaluation module is connected to the barcode repair and recognition module and the image classification module, and is used to integrate the barcode recognition results and the AI ​​classification label to identify the type of item; The item sorting module is connected to the image evaluation module and is used to generate sorting path instructions according to the item type and link the sorting robot arm to execute them.

[0020] The multi-angle image acquisition module includes acquisition devices arranged above and to the side of the sorting line transport device. The acquisition device is a high-definition industrial camera equipped with a ring-shaped LED fill light with adjustable brightness. It is used to improve the lighting conditions of the cargo image, especially in a dark warehouse environment or when the cargo is highly reflective, to ensure the clarity of the barcode and the appearance of the item.

[0021] Subsequently, basic preprocessing is performed on the image information collected by the high-definition industrial camera. The image information includes barcode images and object appearance images. The preprocessing steps include Gaussian smoothing noise reduction and histogram equalization to enhance contrast.

[0022] The high-definition industrial camera used is the Basler acA1920 series, with a resolution of 1920 × 1200 and a frame rate of 30 FPS. The acquisition mode is fixed-point acquisition (still frame). Images are saved in PNG format, with each pixel value ranging from 0–255. Fixed exposure parameters (such as 20ms exposure time and f / 5.6 aperture) are used during acquisition to mitigate the complex lighting conditions of the sorting and transport lines.

[0023] Goods in the logistics warehouse are transported to the scanning area via a conveyor belt. The conveyor belt speed is set to 0.5 m / s to ensure that the camera has enough time to capture the goods from multiple angles.

[0024] When goods enter the scanning area, industrial cameras installed on both sides of the conveyor line automatically trigger the camera function to capture images of the goods from different angles. At the same time, the ring-shaped LED fill light automatically adjusts the brightness based on the ambient light intensity detected by the light sensor to ensure image clarity.

[0025] Perform basic preprocessing on images captured by high-definition industrial cameras, including Gaussian smoothing noise reduction and histogram equalization to enhance contrast. The Gaussian filter kernel size is recommended to be , standard deviation , and then the contrast is enhanced by histogram equalization.

[0026] The preprocessed image information, including the barcode image and the object appearance image, is then sent to the barcode repair and recognition module and the image classification module respectively.

[0027] A high-performance industrial-grade computer equipped with an Intel Core i7 processor, 32GB of memory, and a 1TB solid-state drive is used to meet the computing and storage requirements of the AI ​​repair algorithm and classification model, ensuring stable operation of the platform. It is connected to a network-attached storage (NAS) device with a 10TB storage capacity to store large amounts of scanned images, repaired barcode information, classification results, and other data, facilitating subsequent query statistics and analysis.

[0028] AI repair algorithms include: Recognition unit, used to identify the blurred or damaged area in the barcode image and determine its position and range; The image repair model is trained using a large number of blurred and damaged barcode images and their corresponding standard clear barcode images. The blurred or damaged barcode images are then repaired through the trained deep learning model to generate repaired clear barcode images, and objects are identified based on the repaired clear barcode images.

[0029] For the barcode repair algorithm, a convolutional neural network (CNN) model was trained by collecting 100,000 barcode image samples of various types (such as product barcodes and express delivery number barcodes) with varying degrees of blur and damage, as well as corresponding standard clear barcode images. During training, a learning rate of 0.001 was set, the training cycle was 100 epochs, and a cross-entropy loss function was used. By continuously adjusting the model parameters, the similarity between the repaired barcode images and the standard clear barcode images reached over 95%.

[0030] The AI ​​object image classification model uses transfer learning. Based on a pre-trained ResNet-50 model, it is fine-tuned using 50,000 labeled images of goods in the logistics warehouse (over 20 categories, including electronics, clothing, and food). During training, a learning rate of 0.0001, a training cycle of 50 epochs, and the Adam optimizer were used, achieving a classification accuracy of over 90%.

[0031] Finally, the image restoration accuracy index is introduced to evaluate the restoration results. The calculation formula of the image restoration accuracy index is as follows: , in, represents the image restoration accuracy index, represents the coefficient of variation, represents the ground-corrected image with the coordinates (i, j) on the image, Represents the output image at position (i, j) on the image, Indicates Centered along its main migration direction The extended linear neighborhood window, the greater the degree of deviation, the greater the fluctuation of the shadow of the point in its direction neighborhood, which does not conform to the principle of diffusion path stability. N represents the number of samples, the image restoration accuracy index The closer the value is to 0, the higher the repair accuracy is.

[0032] This formula is the image-level physical stability residual function, which is inspired by the rate gradient consistency principle derived from the two-dimensional simplified diffusion equation, that is, in the main diffusion direction of the image defect or shadow, the velocity change should be smooth, and there should be no sudden changes between the velocity vectors of each point.

[0033] The design of this formula fully considers the particularity of the image scene. The image incompleteness or shadow should not show a sudden change or return, especially under the guidance of the missing direction, the shadow should be consistent in direction and stable in speed. By constructing the direction neighborhood Rather than ordinary region, by enhancing the response to the stationarity in the “missing migration main direction”, thus avoiding being misled by lateral noise interference.

[0034] For example, if the speed of a shadow track between frames is approximately 1.0, 1.1, and 1.2 (pixels / frame), then Close to 0; if the path suddenly changes from 1.0 to 3.5 and then back to 1.2, then The system can judge that the path is "non-stationary diffusion" and its credibility is reduced.

[0035] Another channel of object appearance image data is transmitted to the image classification module. The image classification software first performs feature extraction on the image, using the ResNet-50 model to extract key features such as the object's shape, color, and texture.

[0036] The extracted feature vectors are input into the trained AI classification model for item images. The model classifies and identifies items based on the learned knowledge of the types of goods and assigns corresponding labels for the types of goods, such as "electronic products - mobile phones" and "clothing - T-shirts".

[0037] The classification results are associated and bound with the corresponding barcode information to form complete cargo identification data, which is stored in the storage device and sent to the sorting control system.

[0038] The AI ​​classification model for object images includes: Feature extraction unit, used to extract key features from the object appearance image, including shape, color and texture; Classification training unit, including deep learning classification models, is trained using a large number of labeled item images to accurately identify different types of goods; The classification execution unit is used to classify and identify the newly scanned appearance image of the item based on the trained deep learning classification model, automatically mark the type of goods, generate classification result information, and store it in association with the corresponding barcode information.

[0039] After the camera captures an image of the goods, the image data is first transmitted to the barcode repair and recognition module. This module uses professional barcode recognition software to locate and decode the barcode in the image and outputs a confidence score for the barcode recognition. The confidence score reflects the degree of confidence that the barcode information has been correctly recognized. It is calculated based on multiple characteristics such as the barcode image's clarity, completeness, and match with the standard barcode template.

[0040] For a clear, undamaged barcode that complies with standard encoding rules, the recognition confidence level may be as high as 98%; while for a barcode that is partially covered by tape and has stains, the confidence level may be only 70%.

[0041] Simultaneously, the product image data is also sent to the item classification module. This module uses a trained deep learning classification model to analyze the product's appearance and output a predicted category and corresponding confidence level. The classification confidence level reflects the model's certainty about the classification result and is determined based on factors such as the feature integrity of the product image and its similarity to the training samples. For example, for electronic products with clear images and distinct features, the classification confidence level can reach 95%; for items with blurry appearances and atypical features, the confidence level may drop to 60%.

[0042] When the barcode recognition confidence level exceeds 90%, it indicates a high degree of reliability. In this case, the system prioritizes the use of the barcode's cargo information (such as product number, incoming batch, etc.) for subsequent sorting decisions. This strategy leverages the accuracy and uniqueness of barcode information. When the barcode is reliable, it quickly identifies the cargo, expedites the sorting process, avoids unnecessary complex calculations, and improves system efficiency. For example, a batch of brand-new smartphones just arrived at the warehouse with standardized, clear barcodes and a 95% recognition confidence level. The system immediately determines the cargo type and sorting destination based on the barcode information, eliminating the need to refer to item classification results.

[0043] When the barcode recognition confidence level is less than or equal to 90%, it indicates that there is a certain degree of uncertainty in the barcode information, which may be caused by partial obscuration, wear and tear, or blurred printing. In this case, the system activates the comprehensive decision-making mode, which weights and fuses the item classification results with the barcode information to generate a more accurate cargo identification result.

[0044] The specific weighting method is as follows: , in, represents a weighted decision, is the weighting coefficient, is the barcode recognition result of the item, is the image recognition result of the object.

[0045] and + =1. In the initial state, it can be set based on a large amount of experimental data and experience. =0.6, =0.4, emphasizing the basic role of barcode information while taking into account the supplementary value of item classification.

[0046] Confidence in barcode recognition and item classification confidence Multiply by the corresponding weights and add them together to get the comprehensive confidence ,Right now: , The barcode information is modified or confirmed based on the comprehensive confidence and the item classification results. For example, if the barcode recognition confidence is 80% ( =80%), the item classification model determines that the product is a certain brand of shampoo with a confidence level of 90% ( =90%), calculated according to the above weights =0.6×80%+0.4×90%=84%. If the overall confidence reaches the preset threshold (e.g., 80%), the item classification results are used as auxiliary basis to correct or confirm the category of goods indicated by the barcode information, ensuring the accuracy of the sorting decision. If the overall confidence falls below the threshold, the goods information is marked as questionable, and the manual review process is initiated, with staff making the final decision to ensure foolproof sorting.

[0047] The confidence level C is calculated as follows: , in, Indicates the barcode angle, It represents the attitude angle, which is the counterclockwise rotation angle based on the horizontal line of the image. The unit is degree, and N is the number of samples.

[0048] If the angle between the barcode frame and the barcode shadow changes beyond a certain threshold, it will be considered as unstable perception and the confidence of the frame will be reduced. It automatically drops below the preset threshold, prompting the platform system to temporarily suspend the use of the frame.

[0049] Through this dual-channel analysis and result fusion strategy, the present invention can flexibly and intelligently process cargo information under different conditions, give full play to the respective advantages of barcode recognition and item classification, effectively cope with the complex and changeable cargo identification scenarios in logistics warehouses, further improve sorting efficiency and accuracy, reduce the sorting error rate caused by information errors, and provide solid technical support for the efficient operation of logistics warehouses.

[0050] The article sorting module includes a sorting robot arm, which sorts articles to corresponding target areas according to article types.

[0051] After receiving the cargo identification data from the platform host, the sorting control system plans the sorting routes based on the preset sorting rules and warehouse layout. For example, electronic products can be sorted to the corresponding electronic equipment storage area, and clothing can be sorted to the clothing storage area.

[0052] The sorting control system sends sorting instructions to the sorting robot arm, sorting robot or automatic guided vehicle (AGV), including information such as the location of the goods on the conveyor line, the target storage area, and the transportation route.

[0053] Sorting robotic arms, sorting robots or AGVs accurately grab goods from the conveyor line according to instructions, and transport them to the target storage location along the planned route to complete the sorting and warehousing operations of the goods.

[0054] The image self-correction intelligent scanning platform of the present invention also provides a data update and model optimization mechanism: Regularly collect new barcode types and cargo appearance images in logistics warehouses to expand the training sample set of the barcode repair algorithm and the training data set of the item image classification model, including collecting new data once a quarter, adding 10,000 barcode images and 5,000 item appearance images each time.

[0055] We leveraged the newly added data to retrain and optimize the barcode repair and classification models, continuously improving their generalization and recognition accuracy. This continuous optimization improved the barcode repair model's accuracy for new, blurred, and damaged barcodes, while also increasing the item classification model's accuracy for new types of goods.

[0056] Feedback and error correction mechanisms are also set up: During the goods sorting process, if the sorting robot or staff finds a sorting error (such as incorrect cargo type marking, incorrect barcode information recognition causing the goods to be sent to the wrong storage area, etc.), the error information can be promptly fed back to the image self-correction intelligent scanning platform through a handheld terminal device.

[0057] After receiving the feedback, the platform analyzes the corresponding barcode repair algorithm and image classification model to identify the causes of the errors, such as model parameter deviations and missing barcode damage features. To address these issues, the model is adjusted and optimized, and the relevant model parameters are retrained.

[0058] The optimized model was put back into operation, and the cargo image data that had previously contained errors was reprocessed and recognized to verify the model correction effect, ensure the accuracy of subsequent sorting operations, and continuously improve the stability and reliability of the system.

[0059] Through the above embodiments, the present invention can effectively solve the problem of identifying tape-covered and soiled barcodes in logistics warehouses, reduce the labor cost of sorting centers, and achieve efficient and accurate logistics sorting optimization. It has significant economic and social benefits and can be widely used in various logistics and warehousing companies to promote the intelligent development of the logistics industry.

[0060] See also Figure 2 As shown, in the second aspect of the present invention, a sorting optimization method is also provided, comprising the following steps: S1: High-definition industrial cameras installed above and on the sides of the sorting line transport device dynamically capture multi-view images of items in transport; S2: Evaluate the integrity of the barcode area. If it is blurred or damaged, the barcode will be repaired and recognized using the barcode AI repair algorithm. S3: Parallel execution of dual-channel analysis of barcode recognition and object image classification, where the object image classification is achieved using an image AI classification model; S4: Combine barcode recognition results and AI classification labels to identify the item type; S5: Map the recognition results to the sorting instruction set to drive the sorting robot arm to automatically sort the items.

[0061] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering technicians in this field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. An intelligent scanning platform with self-correction of images, characterized in that: include: A multi-angle image acquisition module is configured above and to the side of the sorting line transport device, and is used to dynamically acquire image information of the items to be identified, including barcode images and images of the item's appearance; A barcode repair and recognition module is connected to the multi-angle image acquisition module. The barcode repair module has a built-in barcode AI repair algorithm for repairing the collected blurred or damaged barcode images and identifying the items based on the repaired barcodes. An image classification module, connected to the multi-angle image acquisition module, having a built-in AI classification model for item images, for automatically marking the type of goods based on the barcode image and the item appearance image during the scanning process; An image evaluation module is connected to the barcode repair and recognition module and the image classification module, and is used to integrate the barcode recognition results and the AI ​​classification label to identify the type of item; The item sorting module is connected to the image evaluation module and is used to generate sorting path instructions according to the item type and link the sorting robot arm to execute them.

2. The image self-correction intelligent scanning platform according to claim 1, characterized in that: The multi-angle image acquisition module includes an acquisition device configured above and on the side of the sorting line transport device. The acquisition device is a high-definition industrial camera. The image information collected by the high-definition industrial camera is then subjected to basic preprocessing. The preprocessing steps include Gaussian smoothing noise reduction and histogram equalization to enhance contrast.

3. The image self-correction intelligent scanning platform according to claim 1, characterized in that: The AI ​​repair algorithm includes: Recognition unit, used to identify the blurred or damaged area in the barcode image and determine its position and range; The image repair model is trained using a large number of blurred and damaged barcode images and their corresponding standard clear barcode images. The blurred or damaged barcode images are then repaired through the trained deep learning model to generate repaired clear barcode images, and objects are identified based on the repaired clear barcode images.

4. The image self-correction intelligent scanning platform according to claim 3, characterized in that: The image restoration accuracy index is introduced to evaluate the restoration results. The calculation formula of the image restoration accuracy index is as follows: , in, represents the image restoration accuracy index, represents the coefficient of variation, represents the ground-corrected image with the coordinates (i, j) on the image, Represents the output image at position (i, j) on the image, Indicates Centered along its main migration direction The extended linear neighborhood window, the greater the degree of deviation, the greater the fluctuation of the shadow of the point in its direction neighborhood, N represents the number of samples, image restoration accuracy index The closer the value is to 0, the higher the repair accuracy is.

5. The image self-correction intelligent scanning platform according to claim 1, characterized in that: The object image AI classification model includes: Feature extraction unit, used to extract key features from the object appearance image, including shape, color and texture; Classification training unit, including deep learning classification models, is trained using a large number of labeled item images to accurately identify different types of goods; The classification execution unit is used to classify and identify the newly scanned appearance image of the item based on the trained deep learning classification model, automatically mark the type of goods, generate classification result information, and store it in association with the corresponding barcode information.

6. The image self-correction intelligent scanning platform according to claim 1, characterized in that: The image evaluation module recognizes the type of item by: When the barcode recognition confidence is greater than 90%, the barcode information is given priority; When the barcode recognition confidence level is less than or equal to 90%, a weighted decision is made based on the item classification results and barcode information. The weighted decision is calculated by the following formula: ， in, represents a weighted decision, is the weighting coefficient, is the barcode recognition result of the item, is the image recognition result of the object.

7. The image self-correction intelligent scanning platform according to claim 1, characterized in that: The article sorting module includes a sorting robot arm, which sorts articles to corresponding target areas according to article types.

8. A sorting optimization method based on the platform according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1: High-definition industrial cameras installed above and on the sides of the sorting line transport device dynamically capture multi-view images of items in transport; S2: Evaluate the integrity of the barcode area. If it is blurred or damaged, the barcode will be repaired and recognized using the barcode AI repair algorithm. S3: Parallel execution of dual-channel analysis of barcode recognition and object image classification, where the object image classification is achieved using an image AI classification model; S4: Combine barcode recognition results and AI classification labels to identify the item type; S5: Map the recognition results to the sorting instruction set to drive the sorting robot arm to automatically sort the items.

Images