Method, system, device and medium for automatically determining pallet defects

Abstract

The present invention relates to a method for automatically determining defects of a pallet, comprising: collecting data on the pallet through an acquisition device to obtain pallet information, analyzing the pallet information to obtain a first image and sensor data; performing defect analysis on the first image through a preset convolutional neural model to obtain first defect information, marking the first defect information on the first image to obtain a marked image; extracting features from the sensor data to obtain sensor feature data, performing defect analysis on the sensor feature data through a preset feature algorithm to obtain second defect information; analyzing the marked image according to the second defect information to obtain third defect information, and marking defects on the corresponding pallet according to the third defect information. The present invention can improve the accuracy and efficiency of detection, reduce the need for manual operation, reduce human errors, and save manpower and time costs.

Description

Technical Field

[0001] The present invention relates to the technical field of determining a defect of a pallet, and in particular to a method, system, device and medium for automatically determining a defect of a pallet. Background Art

[0002] With the continuous improvement of industrial automation, the method of automatically determining pallet defects has become increasingly important in the production field. However, the traditional method of manually detecting pallet defects has problems such as low efficiency, low accuracy and high cost. Summary of the invention

[0003] The main purpose of the present invention is to provide a method, system, device and medium for automatically determining pallet defects, which can improve the accuracy and efficiency of detection, reduce the need for manual operation, reduce human errors, and save manpower and time costs.

[0004] To achieve the above object, the present invention provides a method for automatically determining pallet defects, comprising:

[0005] Collecting data on the pallet through a collection device to obtain pallet information, analyzing the pallet information to obtain a first image and sensor data;

[0006] Performing defect analysis on the first image by using a preset convolutional neural model to obtain first defect information, and marking the first defect information on the first image to obtain a marked image;

[0007] Extracting features from the sensor data to obtain sensor feature data, and performing defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information;

[0008] The marked image is analyzed according to the second defect information to obtain third defect information, and a defect mark is performed on the corresponding pallet according to the third defect information.

[0009] Furthermore, the analyzing the tray information to obtain the first image and sensor data includes:

[0010] Performing data conversion on the pallet information according to a preset data conversion library to obtain pallet data, performing data cleaning processing on the pallet data according to a preset cleaning rule, and performing data extraction on the pallet data after data cleaning processing according to a preset category extraction rule to obtain an initial first image and initial sensor data;

[0011] Determine whether the resolution of the initial first image meets the threshold range, and when it meets the threshold range, set the initial first image as the first image;

[0012] It is determined whether the initial sensing data has abnormal values. If the initial sensing data has abnormal values, the abnormal values ​​are removed from the initial sensing data to obtain sensing data.

[0013] Furthermore, the performing defect analysis on the first image by using a preset convolutional neural model to obtain first defect information includes:

[0014] Performing image feature extraction on the first image by using the convolutional neural model to obtain texture feature data, shape feature data and color feature data, determining whether the texture feature data, shape feature data and color feature data meet preset feature requirements, and marking the first image as defect-free when all image features meet the feature requirements;

[0015] When there are image features that do not meet the feature requirements, the quantity is calculated based on the texture feature data and the shape feature data to obtain the number of defects, the position is calculated based on the texture feature data and the shape feature data to obtain specific defect position data, the range is calculated based on the shape feature data and the color feature data to obtain specific defect range data, the defect quantity, defect position data and defect range data are combined and processed to obtain the first defect information.

[0016] Further, feature extraction is performed on the sensor data to obtain sensor feature data, and defect analysis is performed on the sensor feature data using a preset feature algorithm to obtain second defect information, including:

[0017] By performing feature extraction on the sensor data, vibration sensor feature data, weight sensor feature data, curvature sensor feature data and inclination sensor feature data are obtained; the damage degree of the vibration sensor feature data and the weight sensor feature data is calculated by a feature algorithm to obtain damaged data; structural calculation is performed on the curvature sensor feature data and the inclination sensor feature data by a feature algorithm to obtain structural integrity data; it is respectively determined whether the damaged data and the structural integrity data meet the preset data interval requirements; when they meet the preset data interval requirements, the damaged data and the structural integrity data are combined and converted to obtain second defect information.

[0018] Further, the analyzing the marked image according to the second defect information to obtain third defect information includes:

[0019] Analyze the marked image to obtain a first defect position on the first image, mark the second defect information on the marked image and analyze it to obtain a second defect position on the first image, calculate the deviation between the first defect position and the second defect position to obtain deviation data, determine whether the deviation data meets a preset deviation value interval, and when the deviation value interval is met, perform comprehensive analysis and processing on the first defect information combined with the second defect information to obtain comprehensive defect information, and perform packaging processing on the comprehensive defect information and the marked image to obtain third defect information;

[0020] When the deviation value interval is not met, the marked image is corrected and optimized according to the deviation data to obtain a corrected image, the first defect information and the second defect information are respectively corrected and optimized according to the deviation data to obtain first optimization information and second optimization information, the first optimization information and the second optimization information are comprehensively analyzed and processed to obtain comprehensive optimization information, the corrected image and the comprehensive optimization information are combined and processed to obtain third defect information.

[0021] Further, the step of marking a defect on the corresponding pallet information according to the third defect information includes:

[0022] Read the third defect information to obtain a corresponding pallet number, match the corresponding pallet information according to the pallet number, and perform data association between the third defect information and the pallet information according to a preset association rule;

[0023] When the association is successful, the corresponding pallet is matched through the pallet information, and a defect mark is performed on the pallet according to the third defect information.

[0024] The present invention also provides a system for automatically determining pallet defects, comprising:

[0025] A collection module, the collection module is used to collect data on the pallet through a collection device to obtain pallet information, analyze the pallet information, and obtain a first image and sensor data;

[0026] An analysis module, the analysis module is used to perform defect analysis on the first image through a preset convolutional neural model to obtain first defect information, and mark the first defect information on the first image to obtain a marked image;

[0027] An extraction module, the extraction module is used to perform feature extraction on the sensor data to obtain sensor feature data, and perform defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information;

[0028] A processing module is used to perform correction analysis on the marked image according to the second defect information to obtain third defect information, and perform defect marking on the corresponding pallet according to the third defect information.

[0029] The present invention also provides a device for automatically determining defects of a pallet, comprising:

[0030] Memory, used to store programs;

[0031] The processor is used to execute the program to implement each step of the method for automatically determining pallet defects as described in any one of the above.

[0032] The present invention also provides a storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute any of the above methods.

[0033] The method, system, device and medium for automatically determining pallet defects provided by the present invention have the following beneficial effects:

[0034] Data acquisition equipment is used to collect and analyze data on pallets, and combined with image and sensor data for comprehensive analysis, to achieve automatic identification and marking of pallet defects, improve the accuracy and efficiency of detection, reduce the need for manual operation, reduce human errors, and save manpower and time costs. At the same time, the use of preset models and algorithms such as convolutional neural networks to process and analyze image and sensor data provides a scientific and accurate basis for pallet defect identification, which can more comprehensively and accurately determine the defects of pallets, further improving the accuracy and reliability of identification. The overall process covers multiple links such as data acquisition, image processing, and data analysis, making the detection process more comprehensive, automated, and efficient, and providing convenient and reliable technical support for pallet quality management in industrial production. Through automated defect identification and marking, real-time feedback on the quality status of pallets is provided, prompting timely processing and repair of defects, ensuring the safety and stability of pallets in logistics and transportation, and improving overall production efficiency and product quality levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The present invention provides a flow chart of a method for automatically determining a defect of a pallet;

[0036] Figure 2 The present invention provides a system structure diagram for automatically determining pallet defects;

[0037] Figure 3 The present invention provides a structural diagram of a device for automatically determining defects of a pallet.

[0038] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0040] The present invention is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0041] Reference Figure 1 As shown, the present invention provides a method for automatically determining a pallet defect, comprising:

[0042] Step S1: collecting data on the pallet through a collection device to obtain pallet information, analyzing the pallet information to obtain a first image and sensor data;

[0043] Step S2: performing defect analysis on the first image by using a preset convolutional neural model to obtain first defect information, and marking the first defect information on the first image to obtain a marked image;

[0044] Step S3: extracting features from the sensor data to obtain sensor feature data, and performing defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information;

[0045] Step S4: Analyze the marked image according to the second defect information to obtain third defect information, and mark the defect on the corresponding pallet according to the third defect information.

[0046] Based on the above steps, the specific detailed steps are as follows:

[0047] Step S1: The data acquisition device starts scanning and acquiring the pallet, and obtains the image information and sensor data of the pallet. The acquired data is processed and analyzed, and the pallet information is extracted according to the data features. The acquired image data is processed, including denoising, image enhancement and other operations, to obtain a clear first image.

[0048] Extract and process the collected sensor data to obtain effective information of the sensor data, combine the processed image data and sensor data to generate the first image and sensor data for subsequent defect analysis. Record and save the acquired first image and sensor data for subsequent defect analysis and processing.

[0049] Step S2: Load the pre-trained convolutional neural network model for defect analysis of the first image, and input the first image obtained in the convolutional neural model for analysis. Preprocess the input first image, such as resizing, normalizing, etc., to meet the input requirements of the convolutional neural network. Use the convolutional neural model to perform defect analysis on the first image and detect possible defect areas in the image. According to the model analysis results, the first defect information, i.e., the identified defect location, type, etc., is obtained.

[0050] The first defect information is marked on the first image, and the defect area is marked by using different marking methods for subsequent analysis and display. The marked first image is called a marked image, which is the image result of marking the defect information on the first image, and the marked image with the first defect information is output for subsequent comprehensive analysis and processing.

[0051] Step S3: Extract features from the acquired sensor data. The extraction process includes data cleaning, preprocessing and other steps. After feature extraction, feature information of the sensor data is obtained. In this step, the sensor feature data can be further processed and prepared for defect analysis.

[0052] Load a pre-designed feature algorithm for defect analysis of sensor feature data. Input the processed sensor feature data into the preset feature algorithm for analysis. Use the loaded feature algorithm to perform defect analysis on the sensor feature data. The feature algorithm identifies potential defect information by using machine learning, data mining and other technologies. According to the analysis results of the feature algorithm, obtain the second defect information, that is, the defect content analyzed from the sensor data, such as abnormal state, abnormal data, etc.

[0053] Step S4: Load the marked image containing the first defect information into the system for further analysis. According to the second defect information, analyze the marked image to identify and locate the possible defect area. According to the analysis result of the second defect information, obtain the third defect information, including the specific type, location and possible impact of the defect. According to the third defect information, mark the defect area on the marked image and mark it accordingly, such as marking it with different colors or shapes.

[0054] Map the defect information marked on the marked image to the corresponding actual pallet to determine the specific location and range of the defect on the pallet. According to the defect information on the marked image, mark the defect on the actual pallet, which can be marked by labels, symbols, etc. When marking defects on the pallet, ensure the accuracy and clarity of the mark to facilitate subsequent defect repair work.

[0055] Record the third defect information and defect markings on the pallet, summarize this information, and provide a reference for the formulation of subsequent defect repair plans.

[0056] The method for automatically determining pallet defects provided by the present invention collects and analyzes data on the pallet through a data acquisition device, and performs comprehensive analysis in combination with image and sensor data to realize automatic identification and marking of pallet defects, thereby improving the accuracy and efficiency of detection, reducing the need for manual operation, reducing human errors, and saving manpower and time costs. At the same time, the image and sensor data are processed and analyzed using preset models and algorithms such as convolutional neural networks, providing a scientific and accurate basis for pallet defect identification, and being able to more comprehensively and accurately determine the defect conditions of the pallet, further improving the accuracy and reliability of identification. The overall process covers multiple links such as data acquisition, image processing, and data analysis, making the detection process more comprehensive, automated, and efficient, and providing convenient and reliable technical support for pallet quality management in industrial production. Through automated defect identification and marking, the quality status of the pallet is fed back in real time, prompting timely processing and repair of defects, ensuring the safety and stability of the pallet in logistics and transportation, and improving the overall production efficiency and product quality level.

[0057] In one embodiment, analyzing the tray information to obtain the first image and sensor data includes:

[0058] Use the preset data conversion library to convert the format and standardize the raw pallet information to obtain pallet data to meet the needs of subsequent analysis processes. The data conversion library contains a series of predefined functions or algorithms for processing different types of data (such as text, images, sensor signals, etc.) and converting raw pallet information into a unified data format (such as digital matrices, JSON objects, database records, etc.).

[0059] The converted pallet data is quality checked and error corrected according to the preset cleaning rules. Data cleaning includes removing duplicate values, filling missing values, correcting logical errors, removing abnormal values ​​(such as values ​​outside the normal range), standardizing data formats, etc., to improve data integrity and consistency. Cleaning rules can be customized according to business needs and data characteristics to ensure that the cleaned data can truly reflect the status of the pallet.

[0060] According to the preset category extraction rules, features related to pallet defect detection are extracted from the pallet data after cleaning to generate the initial first image and initial sensor data. The category extraction rules include defining the extraction methods, calculation formulas, threshold settings and other details of various types of features. For example, visual features such as edges, textures, and color distributions are extracted from image data, and physical quantities such as weight, temperature, and vibration frequency are extracted from sensor data.

[0061] The resolution of the initial first image is checked to determine whether it falls within a preset threshold range. The threshold range is usually determined based on factors such as the accuracy requirements of the detection task and the performance of the hardware device. When the resolution of the initial first image meets the threshold range, it is confirmed as the final first image. When it does not meet the threshold range, the image is re-captured or the image processing parameters are adjusted.

[0062] Perform outlier detection on the initial sensor data. When there are outliers, identify the initial sensor data and remove the abnormal data points caused by equipment failure, environmental interference, acquisition error, etc. Common outlier detection methods include statistical methods (such as the three-sigma principle, box plot method), machine learning methods (such as isolation forest, DBSCAN clustering), etc. After removing the outliers, the final sensor data is obtained.

[0063] This embodiment realizes the automatic processing and conversion of pallet information through the preset data conversion library, cleaning rules and category extraction rules, accelerates the speed and accuracy of data processing, and improves efficiency. The preset cleaning rules and category extraction rules can effectively clean and extract pallet data, ensure data quality, and improve the accuracy of subsequent analysis. By judging whether the resolution of the initial first image meets the threshold range and performing corresponding processing, the image quality is optimized and the effect and accuracy of visual analysis are improved. Judging and clearing outliers in sensor data helps to improve the accuracy and stability of sensor data and obtain more reliable data results in subsequent defect analysis.

[0064] In one embodiment, performing defect analysis on the first image by using a preset convolutional neural model to obtain first defect information includes:

[0065] Image features are extracted from the first image through a convolutional neural model to obtain texture feature data, shape feature data and color feature data, and it is determined whether the texture feature data, shape feature data and color feature data meet preset feature requirements. When all image features meet the feature requirements, the first image is marked as defect-free.

[0066] When there are image features that do not meet the feature requirements, the number of defects is determined by quantitative calculation based on the texture feature data and the shape feature data. The calculation process involves clustering, counting or threshold-based segmentation operations on the feature vectors to identify independent defect areas or feature clusters.

[0067] Based on the identified defect-related features, the specific position of each defect is further calculated by locating the maximum response point of the feature vector, calculating the feature center, applying image segmentation algorithms (such as semantic segmentation, instance segmentation), etc., so as to obtain the defect position data of each defect, wherein the defect position data includes the pixel coordinates in the image and the relative position relative to the origin of the pallet.

[0068] The shape feature data and color feature data are used to perform range calculation to determine the defect range data of each defect. This is achieved by calculating the norm of the feature vector, counting the number of continuous pixels, applying the boundary detection algorithm, etc., to obtain the defect range data of each defect, including quantitative indicators such as the defect diameter, circumference, and area.

[0069] The defect quantity, defect location data and defect range data are combined and processed to obtain first defect information.

[0070] This embodiment uses a preset convolutional neural model to perform defect analysis on the first image, and combined with texture, shape and color feature extraction, it can accurately detect and identify defects in the image, thereby improving the accuracy of defect detection. According to the calculation of the number, location and range of defects, the specific location and range of defects can be quickly and accurately located, providing important positioning information for subsequent repair work. Through preset feature requirements and combined processing, it can automatically infer whether there are defects in the image and provide defect information, reducing the need for manual intervention and improving the degree of automation.

[0071] In one embodiment, feature extraction is performed on the sensor data to obtain sensor feature data, and defect analysis is performed on the sensor feature data using a preset feature algorithm to obtain second defect information, including:

[0072] By extracting features from the sensor data, vibration sensor feature data, weight sensor feature data, curvature sensor feature data and inclination sensor feature data are obtained, wherein the vibration sensor feature data include vibration frequency, amplitude, acceleration peak and vibration power spectrum density; the weight sensor feature data include: key weight parameters, such as real-time weight, weight change trend and weight fluctuation amplitude; the curvature sensor feature data include: maximum curvature value, average curvature and curvature distribution characteristics; the inclination sensor feature data include: inclination angle, inclination direction and inclination change rate.

[0073] The damage degree of the vibration sensor feature data and the weight sensor feature data is calculated by the feature algorithm to obtain the damage data. The structure of the bending sensor feature data and the inclination sensor feature data is calculated by the feature algorithm to obtain the structure integrity data. The structure integrity data includes indicators such as the structure stability score, deformation risk level and posture balance state.

[0074] It is determined whether the damaged data and the structural integrity data meet the preset data interval respectively. When the preset data interval requirement is met, the damaged data and the structural integrity data are combined and converted to obtain the second defect information.

[0075] This embodiment can comprehensively analyze the sensor data and discover potential defect problems by extracting and calculating features of the sensor data. By performing defect analysis on the sensor feature data through a preset feature algorithm, the degree of damage and structural integrity can be accurately calculated, thereby improving the accuracy and reliability of defect detection. By calculating and judging the data interval, it is possible to quickly determine whether the sensor data meets the preset requirements, thereby quickly identifying and locating problems and defects in the sensor. The sensor data is automatically analyzed and processed using a feature algorithm to improve the efficiency of data processing, reduce manual intervention, and speed up defect analysis. By combining the conversion of damaged data and structural integrity data, the second defect information is obtained, which can provide more refined and detailed defect information, facilitating subsequent repair and processing. By timely analyzing and processing defects in sensor data, it is helpful to improve the quality control and management level of the production process, and avoid product quality degradation or production failures due to sensor problems.

[0076] In one embodiment, analyzing the marked image according to the second defect information to obtain the third defect information includes:

[0077] Analyze the marked image to obtain the first defect position on the first image, mark the second defect information on the marked image and analyze it to obtain the second defect position on the first image, calculate the deviation between the first defect position and the second defect position to obtain deviation data, determine whether the deviation data meets the preset deviation value interval, and when it meets the deviation value interval, perform comprehensive analysis and processing on the first defect information and the second defect information to obtain comprehensive defect information, and perform packaging processing on the comprehensive defect information and the marked image to obtain third defect information;

[0078] When the deviation value interval is not met, the marked image is corrected and optimized according to the deviation data to obtain a corrected image, the first defect information and the second defect information are corrected and optimized respectively according to the deviation data to obtain first optimization information and second optimization information, the first optimization information and the second optimization information are comprehensively analyzed and processed to obtain comprehensive optimization information, the corrected image and the comprehensive optimization information are combined and processed to obtain third defect information.

[0079] This embodiment can accurately confirm the position of the first defect by analyzing the marked image, combining the second defect information and deviation calculation, thereby improving the accuracy of defect positioning. By comprehensively analyzing and processing the first defect information and the second defect information, comprehensive defect information is obtained, which helps to fully understand the defect situation and the overall problem, and provides an important basis for subsequent processing. When the data does not conform to the deviation value interval, correction and optimization processing is performed to improve the efficiency of defect processing and repair, and ensure that the relevant data meets the preset requirements. Through the comprehensive processing of the first defect information and the second defect information, the accuracy and comprehensiveness of the defect information can be improved, which helps to improve the accuracy of the final defect analysis results. Combining the comprehensive defect information and the marked image for packaging processing can better display the defect information, clearly display and analyze the relationship between the defects, and facilitate subsequent processing and management.

[0080] In one embodiment, marking a defect on the corresponding pallet information according to the third defect information includes:

[0081] Read the third defect information, obtain the corresponding pallet number, match the corresponding pallet information according to the pallet number, and perform data association between the third defect information and the pallet information according to a preset association rule;

[0082] When the association is successful, the corresponding pallet is matched through the pallet information, and a defect mark is made on the pallet according to the third defect information.

[0083] This embodiment ensures accurate defect marking on the correct pallet by associating and matching the third defect information with the corresponding pallet information, thereby improving the accuracy and reliability of marking. By means of automated association and matching, defect information can be quickly processed, and the corresponding pallet can be quickly found and marked, thus shortening the entire processing cycle. Associating the third defect information with the pallet information and marking it can effectively record the defect information and processing process, thus improving data traceability and management convenience.

[0084] Reference Figure 2 As shown, the present invention also provides a system for automatically determining pallet defects, comprising:

[0085] A collection module, the collection module is used to collect data on the pallet through a collection device to obtain pallet information, analyze the pallet information, and obtain a first image and sensor data;

[0086] An analysis module, the analysis module is used to perform defect analysis on the first image through a preset convolutional neural model to obtain first defect information, and mark the first defect information on the first image to obtain a marked image;

[0087] An extraction module is used to extract features from the sensor data to obtain sensor feature data, and to perform defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information;

[0088] The processing module is used to correct and analyze the marked image according to the second defect information to obtain the third defect information, and mark the defect on the corresponding pallet according to the third defect information.

[0089] The system for automatically determining pallet defects provided by the present invention collects and analyzes data on pallets through data acquisition equipment, combines images and sensor data for comprehensive analysis, and realizes automatic identification and marking of pallet defects, thereby improving the accuracy and efficiency of detection, reducing the need for manual operation, reducing human errors, and saving manpower and time costs. At the same time, the image and sensor data are processed and analyzed using preset models and algorithms such as convolutional neural networks, providing a scientific and accurate basis for pallet defect identification, and being able to more comprehensively and accurately determine the defects of pallets, further improving the accuracy and reliability of identification. The overall process covers multiple links such as data acquisition, image processing, and data analysis, making the detection process more comprehensive, automated, and efficient, and providing convenient and reliable technical support for pallet quality management in industrial production. Through automated defect identification and marking, the quality status of the pallet is fed back in real time, prompting timely processing and repair of defects, ensuring the safety and stability of the pallet in logistics and transportation, and improving the overall production efficiency and product quality level.

[0090] Reference Figure 3 As shown, the present invention also provides a device for automatically determining a defect of a pallet, comprising:

[0091] Memory, used to store programs;

[0092] The processor is used to execute the program to implement the various steps of the method for automatically determining pallet defects as described in any one of the above.

[0093] In this embodiment, the processor and the memory may be connected via a bus or other means. The memory may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a read-only memory, a flash memory, a hard disk, or a solid-state drive. The processor may be a general-purpose processor, such as a central processing unit, a digital signal processor, an application-specific integrated circuit, or one or more integrated circuits configured to implement the embodiments of the present invention.

[0094] The present invention also provides a storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute any of the above methods.

[0095] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and conciseness of description, the specific working process of the system and each module described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0096] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for automatically determining pallet defects, characterized in that: include: Collecting data on the pallet through a collection device to obtain pallet information, analyzing the pallet information to obtain a first image and sensor data; Performing defect analysis on the first image by using a preset convolutional neural model to obtain first defect information, and marking the first defect information on the first image to obtain a marked image; Extracting features from the sensor data to obtain sensor feature data, and performing defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information; Analyzing the marked image according to the second defect information to obtain third defect information, and marking a defect on the corresponding pallet according to the third defect information; Extracting features from the sensor data to obtain sensor feature data, and performing defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information, including: By performing feature extraction on the sensor data, vibration sensor feature data, weight sensor feature data, curvature sensor feature data and inclination sensor feature data are obtained; damage degree calculation is performed on the vibration sensor feature data and the weight sensor feature data through a feature algorithm to obtain damaged data; structural calculation is performed on the curvature sensor feature data and the inclination sensor feature data through a feature algorithm to obtain structural integrity data; whether the damaged data and the structural integrity data meet the preset data interval is determined respectively; when the preset data interval requirement is met, the damaged data and the structural integrity data are combined and converted to obtain second defect information; The step of analyzing the marked image according to the second defect information to obtain third defect information includes: Analyze the marked image to obtain a first defect position on the first image, mark the second defect information on the marked image and analyze it to obtain a second defect position on the first image, calculate the deviation between the first defect position and the second defect position to obtain deviation data, determine whether the deviation data meets a preset deviation value interval, and when the deviation value interval is met, perform comprehensive analysis and processing on the first defect information combined with the second defect information to obtain comprehensive defect information, and perform packaging processing on the comprehensive defect information and the marked image to obtain third defect information; When the deviation value interval is not met, the marked image is corrected and optimized according to the deviation data to obtain a corrected image, the first defect information and the second defect information are respectively corrected and optimized according to the deviation data to obtain first optimization information and second optimization information, the first optimization information and the second optimization information are comprehensively analyzed and processed to obtain comprehensive optimization information, the corrected image and the comprehensive optimization information are combined and processed to obtain third defect information.

2. The method for automatically determining pallet defects according to claim 1, characterized in that: The analyzing the tray information to obtain the first image and sensor data includes: Performing data conversion on the pallet information according to a preset data conversion library to obtain pallet data, performing data cleaning processing on the pallet data according to a preset cleaning rule, and performing data extraction on the pallet data after data cleaning processing according to a preset category extraction rule to obtain an initial first image and initial sensor data; Determine whether the resolution of the initial first image meets the threshold range, and if it meets the threshold range, set the initial first image as the first image; It is determined whether the initial sensing data has abnormal values. If the initial sensing data has abnormal values, the abnormal values ​​are removed from the initial sensing data to obtain sensing data.

3. The method for automatically determining pallet defects according to claim 1, characterized in that: The performing defect analysis on the first image by using a preset convolutional neural model to obtain first defect information includes: Performing image feature extraction on the first image by using the convolutional neural model to obtain texture feature data, shape feature data and color feature data, determining whether the texture feature data, shape feature data and color feature data meet preset feature requirements, and marking the first image as defect-free when all image features meet the feature requirements; When there are image features that do not meet the feature requirements, the quantity is calculated based on the texture feature data and the shape feature data to obtain the number of defects, the position is calculated based on the texture feature data and the shape feature data to obtain specific defect position data, the range is calculated based on the shape feature data and the color feature data to obtain specific defect range data, the defect quantity, defect position data and defect range data are combined and processed to obtain the first defect information.

4. The method for automatically determining pallet defects according to claim 1, characterized in that: The step of marking a defect on the corresponding pallet information according to the third defect information includes: Read the third defect information to obtain a corresponding pallet number, match the corresponding pallet information according to the pallet number, and perform data association between the third defect information and the pallet information according to a preset association rule; When the association is successful, the corresponding pallet is matched through the pallet information, and a defect mark is performed on the pallet according to the third defect information.

5. A system for automatically determining pallet defects, characterized in that: include: A collection module, the collection module is used to collect data on the pallet through a collection device to obtain pallet information, analyze the pallet information, and obtain a first image and sensor data; An analysis module, the analysis module is used to perform defect analysis on the first image through a preset convolutional neural model to obtain first defect information, and mark the first defect information on the first image to obtain a marked image; An extraction module, the extraction module is used to perform feature extraction on the sensor data to obtain sensor feature data, and perform defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information; A processing module, the processing module is used to correct and analyze the marked image according to the second defect information to obtain third defect information, and mark the defect on the corresponding pallet according to the third defect information; Extracting features from the sensor data to obtain sensor feature data, and performing defect analysis on the sensor feature data using a preset feature algorithm to obtain second defect information, including: By performing feature extraction on the sensor data, vibration sensor feature data, weight sensor feature data, curvature sensor feature data and inclination sensor feature data are obtained; damage degree calculation is performed on the vibration sensor feature data and the weight sensor feature data through a feature algorithm to obtain damaged data; structural calculation is performed on the curvature sensor feature data and the inclination sensor feature data through a feature algorithm to obtain structural integrity data; whether the damaged data and the structural integrity data meet the preset data interval is determined respectively; when the preset data interval requirement is met, the damaged data and the structural integrity data are combined and converted to obtain second defect information; The correcting and analyzing the marked image according to the second defect information to obtain third defect information includes: Analyze the marked image to obtain a first defect position on the first image, mark the second defect information on the marked image and analyze it to obtain a second defect position on the first image, calculate the deviation between the first defect position and the second defect position to obtain deviation data, determine whether the deviation data meets a preset deviation value interval, and when the deviation value interval is met, perform comprehensive analysis and processing on the first defect information combined with the second defect information to obtain comprehensive defect information, and perform packaging processing on the comprehensive defect information and the marked image to obtain third defect information; When the deviation value interval is not met, the marked image is corrected and optimized according to the deviation data to obtain a corrected image, the first defect information and the second defect information are respectively corrected and optimized according to the deviation data to obtain first optimization information and second optimization information, the first optimization information and the second optimization information are comprehensively analyzed and processed to obtain comprehensive optimization information, the corrected image and the comprehensive optimization information are combined and processed to obtain third defect information.

6. A device for automatically determining pallet defects, characterized in that: include: Memory, used to store programs; The processor is used to execute the program to implement the various steps of the method for automatically determining pallet defects as described in any one of claims 1 to 4.

7. A storage medium, characterized in that: Computer instructions are stored, and the computer instructions are used to make a computer execute the method according to any one of claims 1 to 4.

Images