Intelligent product quality risk monitoring method and system based on convolutional neural network
Through the intelligent product quality risk monitoring method based on convolutional neural network, the problem of inefficient monitoring of quality risk of intelligent robot dogs is solved, automated monitoring and early warning are realized, and monitoring efficiency and maintenance efficiency are improved.
Patent Information
- Application Number
- CN202410442790.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-12
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-04-12
AI Technical Summary
The existing intelligent robot dog quality risk monitoring system is complex in design, resulting in low monitoring efficiency and difficult maintenance, and inefficient problems.
The intelligent product quality risk monitoring method based on convolutional neural network is adopted, and the quality risk data of intelligent robot dogs is obtained for pre-processing, and the convolutional neural network model is used to judge the appearance damage, motion status and interaction conditions, so as to achieve automated early warning and monitoring.
It improves the efficiency of monitoring the quality risk of intelligent robot dogs, reduces labor costs, promptly detects and warns of damaged or abnormal movements, and improves maintenance efficiency and operating conditions monitoring capabilities.
Smart Images

Figure CN118333389B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intelligent product quality monitoring, and in particular to an intelligent product quality risk monitoring method and system based on a convolutional neural network. Background Art
[0002] With the popularization of the Internet and the rise of the Internet of Things, an interconnected and mutually supportive intelligent system has been formed, which provides a foundation for the development of smart products. Smart products can provide a smarter, more convenient and more comfortable life experience. For example, smart home products can realize remote control, automatic adjustment and other functions, smart wearable devices can monitor health status and sports data, and smart toys can talk and have some simple interactions with people. Due to the increasing importance of smart products in life and work, product quality problems may lead to serious consequences, such as safety hazards, data leakage, and reduced user experience. Therefore, companies and manufacturers are paying more and more attention to product quality management and strengthening the monitoring and control of quality risks. Convolutional neural network is a deep learning model commonly used in image and video processing. Compared with traditional neural networks, convolutional neural network has better ability to process image and sequence data because it can automatically learn features in images and extract the most useful information. In this context, convolutional neural network, as an emerging technology, provides new ideas and methods for quality risk monitoring of smart robot dogs.
[0003] With the development of big data and artificial intelligence technology, enterprises can use data analysis and mining technology to monitor product quality. Smart products can generate a large amount of data, including sensor data, usage data, etc. By analyzing this data, potential quality problems can be discovered in a timely manner and corresponding measures can be taken to deal with them; at the same time, by analyzing and mining this data to provide personalized services, optimize user experience, and even make predictions and decisions. The existing product quality monitoring system obtains product data and data on factors affecting product quality to analyze and determine whether the standards are met.
[0004] For example, a livestock product quality monitoring system disclosed in a patent application with publication number: CN113049776A includes: an air quality monitoring unit, a soil quality monitoring unit, a product quality monitoring unit, a product acquisition unit, a first signal processing unit, a second signal processing unit, a third signal processing unit, a central processing unit, a comparison module one, a comparison module two, a comparison module three and an image display unit; the air quality monitoring unit can be used to monitor the heavy metal content in the soil to prevent cattle and sheep from accidentally ingesting the heavy metals and affecting the meat quality; the product quality monitoring unit can monitor the moisture, lean meat and fat content in the meat, and can accurately check whether the meat quality meets the standards; and the comparison modules one, two and three can also be used to compare the air quality, soil quality and product quality information with normal indicators one by one, and further determine whether the growth environment of cattle and sheep and the meat quality meet the standards.
[0005] For example, the invention patent announcement with announcement number: CN114280076B announces a product surface powder spraying quality monitoring and analysis management system based on machine vision, including: product surface coating appearance defect recognition module, product surface coating detection area division module, spraying database, coating detection area basic parameter acquisition module, coating performance parameter acquisition module, quality modeling and analysis module, supervision cloud center and display terminal. By performing appearance defect detection, basic parameter detection and performance parameter detection on the surface coating of the powder spraying product respectively, the spraying appearance quality coefficient, spraying basic parameter quality coefficient and spraying performance parameter quality coefficient corresponding to the surface coating of the powder spraying product are obtained, so as to comprehensively judge whether the spraying quality of the powder spraying product is qualified.
[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:
[0007] In the prior art, the process of acquiring data includes multiple modules and components, which requires complex hardware and software integration, thereby increasing the difficulty of system design, maintenance and upgrading, and there is a problem of inefficient quality risk monitoring of intelligent robot dogs. Summary of the invention
[0008] The embodiments of the present application solve the problem of inefficient quality risk monitoring of intelligent robot dogs in the prior art by providing an intelligent product quality risk monitoring method and system based on a convolutional neural network, thereby improving the efficiency of quality risk monitoring of intelligent robot dogs and reducing labor costs.
[0009] The embodiment of the present application provides a method for monitoring quality risks of intelligent products based on a convolutional neural network, comprising the following steps: obtaining data and preprocessing: obtaining quality risk data of an intelligent robot dog, the quality risk data including appearance image data, motion performance data, and robot dog instruction data, and performing preprocessing according to the appearance image data, motion performance data, and robot dog instruction data; inspecting the appearance of a robot dog for damage: judging the appearance damage of the intelligent robot dog based on a convolutional neural network model of the appearance image of the intelligent robot dog, and performing early warning processing according to the appearance damage of the intelligent robot dog; inspecting the working status of the robot dog: obtaining a robot dog motion index according to the motion performance data analysis, and performing early warning processing according to the robot dog motion index; inspecting the interaction of the robot dog: obtaining an interaction index of the robot dog according to the robot dog instruction data analysis, and performing early warning processing according to the interaction index of the robot dog.
[0010] Furthermore, the preprocessing according to the appearance image data, the motion performance data, and the robot dog command data is specifically analyzed as follows: extracting the appearance image data, the motion performance data, and the robot dog command data; obtaining appearance image data of uniform size by resizing the appearance image data; performing denoising on the resized appearance image data, the motion performance data, and the robot dog command data to eliminate noise in the image; and performing standardization and normalization on the denoised appearance image data, the motion performance data, and the robot dog command data to obtain appearance image data, the motion performance data, and the robot dog command data of uniform scale and range.
[0011] Furthermore, the robot dog appearance damage inspection is used to analyze the appearance damage of the intelligent robot dog, and the specific analysis is: extracting preprocessed appearance image data; inputting the preprocessed appearance image data into a convolutional neural network model based on the intelligent robot dog appearance image for training; if the convolutional neural network model based on the intelligent robot dog appearance image determines that the appearance of the intelligent robot dog is abnormal, an early warning is issued to notify relevant personnel, and the robot dog working status inspection is not performed; if the convolutional neural network model based on the intelligent robot dog appearance image determines that the appearance of the intelligent robot dog is normal, the robot dog working status inspection is performed.
[0012] Furthermore, the robot dog working status check is used to analyze whether the movement performance of the intelligent robot dog is normal, and the specific analysis is: extracting the pre-processed movement performance data; the movement performance data includes the robot dog walking data and the robot dog turning data; obtaining the robot dog walking index and the robot dog turning index based on the movement performance data analysis; obtaining the robot dog movement index based on the robot dog walking index and the robot dog turning index analysis; judging the working status of the intelligent robot dog based on the robot dog movement index, if the movement performance of the intelligent robot dog is abnormal, an early warning is issued to notify relevant personnel, and the robot dog interaction check is not performed; if the movement performance of the intelligent robot dog is normal, a robot dog interaction check is performed.
[0013] Furthermore, the robot dog walking index is used to analyze the degree of shaking of the intelligent robot dog when walking, and the specific analysis is: extracting the robot dog walking data; the robot dog walking data includes the robot dog leg length, the robot dog joint friction coefficient, and the robot dog walking speed; the robot dog joint friction coefficient is used to describe the flexibility of the intelligent robot dog joints; and obtaining the robot dog walking index based on the robot dog walking data analysis.
[0014] Furthermore, the robot dog turning index is used to describe the smoothness of the intelligent robot dog when turning, and the specific analysis is as follows: extracting the robot dog turning data; the robot dog turning data includes the robot dog turning angle, the robot dog turning speed, the robot dog body length, and the robot dog weight; the robot dog turning angle is used to describe the angle at which the intelligent robot dog executes an instruction to turn its head to drive the body to turn; and obtaining the robot dog turning index based on the analysis of the robot dog turning data.
[0015] Furthermore, the robot dog motion index is used to describe the working condition of the intelligent robot dog, and the specific analysis is: extracting the robot dog walking index and the robot dog turning index; according to the objective weight method, setting the weight ratio of the robot dog walking index to the robot dog motion index and the weight ratio of the robot dog turning index to the robot dog motion index; according to the robot dog walking index, the weight ratio of the robot dog walking index to the robot dog motion index, the robot dog turning index, and the weight ratio of the robot dog turning index to the robot dog motion index, the robot dog motion index is obtained.
[0016] Furthermore, the robot dog interaction check is used to analyze the response of the intelligent robot dog to instructions. The specific process is: extract the pre-processed robot dog instruction data; the robot dog instruction data includes instruction response time, robot dog movement distance, and robot dog turning angle; the instruction response time is used to describe the time interval from the sending of the instruction to the start of execution of the instruction by the intelligent robot dog; the robot dog movement distance is used to describe the distance the intelligent robot dog moves forward or backward when executing the instruction; and the robot dog interaction index is obtained by analyzing the robot dog instruction data.
[0017] Furthermore, the robot dog interaction index is used to describe the accuracy of the intelligent robot dog's response to instructions; the specific formula of the robot dog interaction index is:
[0018]
[0019] Where t represents the command response time, represents the standard value of command response, x represents the movement distance of the robot dog, g represents the steering angle of the robot dog, and J represents the interaction index of the robot dog; obtain the standard value of robot dog interaction; the robot dog interaction standard value is used to describe the minimum accurate value of the intelligent robot dog's response to the command; if the robot dog interaction index is less than the robot dog interaction standard value, the intelligent robot dog fails to execute the command in a timely and accurate manner, and an early warning is issued to notify the relevant personnel; if the robot dog interaction index is not less than the robot dog interaction standard value, the intelligent robot dog executes the command in a timely and accurate manner, and no early warning is issued.
[0020] The present embodiment provides a smart product quality risk monitoring system based on a convolutional neural network, comprising: a data acquisition and preprocessing module, a robot dog appearance damage inspection module, a robot dog working status inspection module, and a robot dog interaction inspection module; wherein the data acquisition and preprocessing module is used to acquire the quality risk data of the smart robot dog, the quality risk data including appearance image data, motion performance data, and robot dog instruction data, and preprocessing is performed according to the appearance image data, motion performance data, and robot dog instruction data; the robot dog appearance damage inspection module is used to judge the appearance damage of the smart robot dog based on a convolutional neural network model of the smart robot dog appearance image, and perform early warning processing according to the appearance damage of the smart robot dog; the robot dog working status inspection module is used to obtain the robot dog motion index according to the motion performance data analysis, and perform early warning processing according to the robot dog motion index; the robot dog interaction inspection module is used to obtain the robot dog interaction index according to the robot dog instruction data analysis, and perform early warning processing according to the robot dog interaction index.
[0021] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:
[0022] 1. By obtaining the quality risk data of the intelligent robot dog, preprocessing is performed according to the quality risk data; the convolutional neural network model based on the appearance image of the intelligent robot dog is used to judge the damage of the appearance of the intelligent robot dog, and early warning processing is performed according to the damage of the appearance of the intelligent robot dog; the robot dog working status inspection is used to describe whether the movement performance of the intelligent robot dog is normal, and the robot dog walking index and the robot dog turning index are obtained according to the movement performance data analysis; the robot dog movement index is obtained according to the robot dog walking index and the robot dog turning index analysis; early warning processing is performed according to the robot dog movement index; the robot dog interaction index is obtained according to the robot dog instruction data analysis, and early warning processing is performed according to the robot dog interaction index, so as to automatically check whether the quality of the intelligent robot dog is normal, thereby realizing the improvement of the efficiency of the quality risk monitoring of the intelligent robot dog, reducing labor costs, and effectively solving the problem of inefficient quality risk monitoring of the intelligent robot dog in the prior art.
[0023] 2. Through automated detection of the appearance and movement performance of the smart robot dog, timely detection and warning of damage or abnormal movement performance of the smart robot dog can be provided, thereby timely discovering possible problems and improving the maintenance efficiency of the smart robot dog.
[0024] 3. By monitoring the status of the intelligent robot dog during various inspections, early warnings can be issued for possible problems in a certain inspection, thereby improving the ability to monitor the operation of the intelligent robot dog. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 A flow chart of a method for intelligent product quality risk monitoring based on a convolutional neural network provided in an embodiment of the present application;
[0026] Figure 2 A schematic diagram of the structure of an intelligent product quality risk monitoring system based on a convolutional neural network provided in an embodiment of the present application. DETAILED DESCRIPTION
[0027] The embodiment of the present application solves the problem of inefficient quality risk monitoring of intelligent robot dogs in the prior art by providing an intelligent product quality risk monitoring method and system based on a convolutional neural network. The method obtains quality risk data of the intelligent robot dog and performs preprocessing based on the quality risk data; judges the appearance damage of the intelligent robot dog based on a convolutional neural network model of the appearance image of the intelligent robot dog, and performs early warning processing based on the appearance damage of the intelligent robot dog; the robot dog working status check is used to describe whether the movement performance of the intelligent robot dog is normal, and the robot dog walking index and the robot dog turning index are obtained based on the movement performance data analysis; the robot dog movement index is obtained based on the robot dog walking index and the robot dog turning index analysis; early warning processing is performed based on the robot dog movement index; the robot dog interaction index is obtained based on the robot dog instruction data analysis, and early warning processing is performed based on the robot dog interaction index, thereby automatically checking whether the quality of the intelligent robot dog is normal, thereby improving the efficiency of quality risk monitoring of the intelligent robot dog and reducing labor costs.
[0028] The technical solution in the embodiment of the present application is to solve the problem of inefficient quality risk monitoring of the above-mentioned intelligent robot dog, and the overall idea is as follows:
[0029] By obtaining the quality risk data of the intelligent robot dog, the quality risk data includes appearance image data, motion performance data, and robot dog instruction data; preprocessing is performed based on the appearance image data, motion performance data, and robot dog instruction data, and the preprocessing includes size adjustment, denoising, standardization, and normalization; the convolutional neural network model based on the appearance image of the intelligent robot dog is used to judge the damage of the appearance of the intelligent robot dog, if the appearance of the intelligent robot dog is damaged, an early warning is issued to notify relevant personnel, if the appearance of the intelligent robot dog is normal, the robot dog working status is checked; the robot dog working status check is used to describe whether the motion performance of the intelligent robot dog is normal, and the preprocessed motion performance data is extracted, and the motion performance data includes the robot dog walking data and the robot dog turning data; the robot dog walking index and the robot dog turning index are obtained according to the motion performance data analysis; the robot dog motion index is obtained according to the robot dog walking index and the robot dog turning index analysis; the judgment is made according to the robot dog motion index, if the motion performance of the intelligent robot dog is abnormal, an early warning is issued to notify relevant personnel, if the motion performance of the intelligent robot dog is normal, the robot dog interaction check is performed; the robot dog interaction index is obtained according to the robot dog instruction data analysis, and an early warning is processed according to the robot dog interaction index.
[0030] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0031] like Figure 1 As shown, it is a flow chart of the intelligent product quality risk monitoring method based on convolutional neural network provided in an embodiment of the present application, the method is applied to the intelligent product quality risk monitoring system based on convolutional neural network, and the method includes the following steps: data acquisition and preprocessing: acquiring the quality risk data of the intelligent robot dog, the quality risk data including appearance image data, motion performance data, and robot dog instruction data, and preprocessing is performed according to the appearance image data, motion performance data, and robot dog instruction data; robot dog appearance damage inspection: judging the appearance damage of the intelligent robot dog based on the convolutional neural network model of the intelligent robot dog appearance image, and performing early warning processing according to the appearance damage of the intelligent robot dog; robot dog working status inspection: obtaining the robot dog motion index according to the motion performance data analysis, and performing early warning processing according to the robot dog motion index; robot dog interaction inspection: obtaining the robot dog interaction index according to the robot dog instruction data analysis, and performing early warning processing according to the robot dog interaction index.
[0032] Furthermore, the preprocessing based on the appearance image data, the motion performance data, and the robot dog command data is specifically analyzed as follows: extracting the appearance image data, the motion performance data, and the robot dog command data; obtaining appearance image data of uniform size by resizing the appearance image data; performing denoising on the resized appearance image data, the motion performance data, and the robot dog command data to eliminate noise in the image; and performing standardization and normalization on the denoised appearance image data, the motion performance data, and the robot dog command data to obtain appearance image data, the motion performance data, and the robot dog command data of uniform scale and range.
[0033] In this embodiment, the appearance image data, the motion performance data, and the robot dog command data are normalized using the Z-score normalization method.
[0034]
[0035]
[0036]
[0037] In the formula, y1 represents the appearance image data, μ1 represents the mean of the appearance image data, σ1 represents the standard deviation of the appearance image data, and y′1 represents the motion performance data after normalization; y2 represents the motion performance data, μ2 represents the mean of the motion performance data, σ2 represents the standard deviation of the motion performance data, and y′2 represents the motion performance data after normalization; y3 represents the robot dog command data, μ3 represents the mean of the robot dog command data, σ3 represents the standard deviation of the robot dog command data, and y′3 represents the robot dog command data after normalization. By resizing the appearance image data to obtain appearance image data of uniform size, it is helpful to use the same processing method for all images in subsequent processing steps to avoid problems that may be caused by inconsistent sizes. Images of uniform size are also helpful to save computing resources and speed up processing. By denoising the resized appearance image data, motion performance data, and robot dog command data, the data quality can be improved, making it easier for the model to capture real patterns and features. By standardizing and normalizing the denoised appearance image data, motion performance data, and robot dog command data, the performance and generalization ability of the model can be improved, while reducing the risk of overfitting.
[0038] Furthermore, the robot dog appearance damage inspection is used to analyze the appearance damage of the intelligent robot dog, and the specific analysis is: extracting the preprocessed appearance image data; inputting the preprocessed appearance image data into the convolutional neural network model based on the intelligent robot dog appearance image for training; if the convolutional neural network model based on the intelligent robot dog appearance image determines that the appearance of the intelligent robot dog is abnormal, an early warning is issued to notify relevant personnel, and the robot dog working status inspection is not performed; if the convolutional neural network model based on the intelligent robot dog appearance image determines that the appearance of the intelligent robot dog is normal, the robot dog working status inspection is performed.
[0039] Furthermore, the robot dog working status check is used to analyze whether the movement performance of the intelligent robot dog is normal. The specific analysis is: extract the pre-processed movement performance data; the movement performance data includes the robot dog walking data and the robot dog turning data; obtain the robot dog walking index and the robot dog turning index based on the movement performance data analysis; obtain the robot dog movement index based on the robot dog walking index and the robot dog turning index analysis; judge the working status of the intelligent robot dog based on the robot dog movement index, if the movement performance of the intelligent robot dog is abnormal, an early warning will be issued to notify relevant personnel, and the robot dog interaction check will not be performed; if the movement performance of the intelligent robot dog is normal, a robot dog interaction check will be performed.
[0040] In this embodiment, if the robot dog motion index is less than 70%, the intelligent robot dog's motion is restricted or subject to a certain degree of difficulty, and the motion is not smooth. An early warning is issued to notify relevant personnel, and the robot dog interaction check is not performed; if the robot dog motion index is not less than 70%, the intelligent robot dog moves relatively smoothly, has good mobility and steering flexibility, and the robot dog interaction check is performed.
[0041] Furthermore, the robot dog walking index is used to analyze the degree of shaking of the intelligent robot dog when walking. The specific analysis is: extracting the robot dog walking data; the robot dog walking data includes the robot dog leg length, the robot dog joint friction coefficient, and the robot dog walking speed; the robot dog joint friction coefficient is used to describe the flexibility of the intelligent robot dog joints; and obtaining the robot dog walking index based on the robot dog walking data analysis.
[0042] In this embodiment, the range of the friction coefficient of the robot dog's joints is 0.01-0.02. In addition to being obtained by simulating the movement of the intelligent robot dog through computer simulation or simulation software, the robot dog's walking index can also be obtained by calculation:
[0043] Z=cosh -1 [sinh -1 (c*v*u)+1]
[0044] In the formula, c represents the leg length of the robot dog, v represents the walking speed of the robot dog, u represents the friction coefficient of the robot dog's joints, and Z represents the walking index of the robot dog. The leg length of the robot dog is directly obtained by measuring tools, such as a ruler and a rangefinder; the walking speed of the robot dog is directly measured by a speed sensor.
[0045] Furthermore, the robot dog turning index is used to describe the smoothness of the intelligent robot dog when turning. The specific analysis is as follows: extract the robot dog turning data; the robot dog turning data includes the robot dog turning angle, the robot dog turning speed, the robot dog body length, and the robot dog weight; the robot dog turning angle is used to describe the angle at which the intelligent robot dog turns its head to drive the body to turn when executing instructions; the robot dog turning index is obtained based on the analysis of the robot dog turning data.
[0046] In this embodiment, the robot dog body length is used to describe the body length of the intelligent robot dog from the head to the buttocks, and the robot dog turning speed is used to describe the speed at which the head drives the body to turn; the robot dog turning index can be obtained by modeling and simulating the turning action of the robot dog through computer simulation or simulation software, and can also be obtained by calculation:
[0047]
[0048] In the formula, g represents the robot dog's steering angle, v' represents the robot dog's steering speed, c' represents the robot dog's body length, m represents the robot dog's weight, and A represents the robot dog's steering index; the robot dog's steering angle multiplied by the robot dog's steering speed shows a negative correlation with the robot dog's steering speed. The larger the robot dog's steering angle, the greater the inertia of the intelligent robot dog's body, and the more likely it is to fall and fail to turn successfully. The faster the robot dog's steering speed, the more likely it is to fall and fail to turn successfully due to the speed difference between the head and the body of the intelligent robot dog. The robot dog's steering angle is obtained through a rotary encoder, an inertial measurement unit, simulation, and emulation; the robot dog's steering speed is obtained through an encoder, a sensor, and an inertial measurement unit; the robot dog's body length is obtained through measuring tools, including rulers and measuring instruments; the robot dog's weight is obtained through weighing equipment, including ordinary weighing scales and professional electronic weighing devices.
[0049] Furthermore, the robot dog motion index is used to describe the working condition of the intelligent robot dog. The specific analysis is as follows: extract the robot dog walking index and the robot dog turning index; according to the objective weight method, set the weight ratio of the robot dog walking index to the robot dog motion index and the weight ratio of the robot dog turning index to the robot dog motion index; according to the robot dog walking index, the weight ratio of the robot dog walking index to the robot dog motion index, the robot dog turning index, and the weight ratio of the robot dog turning index to the robot dog motion index, the robot dog motion index is obtained.
[0050] In this embodiment, the robot dog motion index can be obtained by modeling and simulating the robot dog's motion through computer simulation or simulation software, or by calculation:
[0051]
[0052] In the formula, A represents the robot dog turning index, A′ represents the weight ratio of the robot dog turning index to the robot dog motion index, Z represents the robot dog walking index, Z′ represents the weight ratio of the robot dog walking index to the robot dog motion index, e represents the natural constant, and D represents the robot dog motion index. The robot dog motion index can be used to comprehensively evaluate the performance of intelligent robot dogs in walking and turning, rather than just evaluating a single indicator; comprehensive consideration of different indicators such as walking and turning can help discover the advantages and disadvantages of robot dogs in different aspects and provide a basis for improvement; the objective weight method is used to set the weight ratio value to avoid the influence of subjective factors on the evaluation results, thereby improving the objectivity and accuracy of the evaluation.
[0053] Furthermore, the robot dog interaction check is used to analyze the response of the intelligent robot dog to instructions. The specific process is: extract the pre-processed robot dog instruction data; the robot dog instruction data includes instruction response time, robot dog movement distance, and robot dog turning angle; the instruction response time is used to describe the time interval from the sending of the instruction to the start of execution of the instruction by the intelligent robot dog; the robot dog movement distance is used to describe the distance the intelligent robot dog moves forward or backward when executing the instruction; the robot dog interaction index is obtained by analyzing the robot dog instruction data.
[0054] Furthermore, the robot dog interaction index is used to describe the accuracy of the intelligent robot dog's response to instructions; the specific formula of the robot dog interaction index is:
[0055]
[0056] Where t represents the command response time, represents the standard value of command response, x represents the robot dog's movement distance, g represents the robot dog's turning angle, and J represents the robot dog's interaction index; obtain the robot dog's interaction standard value; the robot dog's interaction standard value is used to describe the lowest accurate value of the intelligent robot dog's response to commands; if the robot dog's interaction index is less than the robot dog's interaction standard value, the intelligent robot dog fails to execute the command in a timely and accurate manner, and an early warning is issued to notify relevant personnel; if the robot dog's interaction index is not less than the robot dog's interaction standard value, the intelligent robot dog executes the command in a timely and accurate manner, and no early warning is issued.
[0057] In this embodiment, the range of the robot dog interaction standard value is 99% to 100%; the instruction response standard value is used to describe the minimum time of 100ms from the issuance of an instruction to the start of execution of the instruction by the intelligent robot dog; in addition to being obtained through calculation, the robot dog interaction index can also be obtained through simulation testing. The specific process is: determine the test scenario, design test cases according to the test scenario, set up the test environment, test the intelligent robot dog according to the test case, collect the data generated during the test, and analyze the test results based on the collected data, so as to evaluate the accuracy of the intelligent robot dog's response to instructions.
[0058] like Figure 2 As shown, it is a structural schematic diagram of the intelligent product quality risk monitoring system based on convolutional neural network provided in the embodiment of the present application. The intelligent product quality risk monitoring system based on convolutional neural network provided in the embodiment of the present application includes: a data acquisition and preprocessing module, a robot dog appearance damage inspection module, a robot dog working status inspection module, and a robot dog interaction inspection module; wherein, the data acquisition and preprocessing module is used to acquire the quality risk data of the intelligent robot dog, and the quality risk data includes appearance image data, motion performance data, and robot dog instruction data, and preprocessing is performed according to the appearance image data, motion performance data, and robot dog instruction data; the robot dog appearance damage inspection module is used to judge the appearance damage of the intelligent robot dog based on the convolutional neural network model of the intelligent robot dog appearance image, and perform early warning processing according to the appearance damage of the intelligent robot dog; the robot dog working status inspection module is used to obtain the robot dog motion index according to the motion performance data analysis, and perform early warning processing according to the robot dog motion index; the robot dog interaction inspection module is used to obtain the robot dog interaction index according to the robot dog instruction data analysis, and perform early warning processing according to the robot dog interaction index.
[0059] The technical solution in the above-mentioned embodiment of the present application has at least the following technical effects or advantages: relative to a livestock product quality monitoring system disclosed in the patent application with publication number: CN113049776A, the embodiment of the present application automatically detects the appearance condition and movement performance of the intelligent robot dog, timely detects and warns of damage or abnormal movement performance of the intelligent robot dog, thereby timely discovers possible problems, thereby improving the maintenance efficiency of the intelligent robot dog; relative to a product surface powder spraying quality monitoring and analysis management system based on machine vision announced in the invention patent with publication number: CN114280076B, the embodiment of the present application monitors the status of the intelligent robot dog during various inspections, thereby warning of possible problems that may arise in a certain inspection, thereby improving the ability to monitor the operation of the intelligent robot dog.
[0060] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0061] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0062] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0063] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0064] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0065] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. An intelligent product quality risk monitoring method based on convolutional neural network, It is characterized in that The following steps are involved: Acquire data and preprocess: acquire quality risk data of the intelligent robot dog, which includes appearance image data, motion performance data, and robot dog instruction data, and preprocess the data according to the appearance image data, motion performance data, and robot dog instruction data; Robot dog appearance damage inspection: The convolutional neural network model based on the appearance image of the intelligent robot dog is used to judge the appearance damage of the intelligent robot dog, and early warning processing is carried out according to the appearance damage of the intelligent robot dog; Robot dog working status check: obtain the robot dog's motion index based on the motion performance data analysis, and conduct early warning processing based on the robot dog's motion index; Robot dog interaction check: The robot dog interaction index is obtained based on the robot dog instruction data analysis, and early warning processing is performed based on the robot dog interaction index; The robot dog working status check is used to analyze whether the movement performance of the intelligent robot dog is normal, and the specific analysis is as follows: extracting pre-processed sports performance data; The motion performance data includes robot dog walking data and robot dog turning data; The robot dog walking index and robot dog turning index are obtained based on the analysis of the sports performance data; The robot dog movement index is obtained by analyzing the robot dog walking index and the robot dog turning index; The working status of the intelligent robot dog is judged according to the robot dog's motion index. If the intelligent robot dog's motion performance is abnormal, an early warning is issued to notify relevant personnel, and the robot dog interaction check is not performed; If the movement performance of the intelligent robot dog is normal, the robot dog interaction check is performed; The robot dog walking index is used to analyze the shaking degree of the intelligent robot dog when walking, and the specific analysis is: Extract the robot dog walking data; The robot dog walking data includes the robot dog's leg length, the robot dog's joint friction coefficient, and the robot dog's walking speed; The friction coefficient of the robot dog joint is used to describe the flexibility of the intelligent robot dog joint; The robot dog walking index is obtained based on the robot dog walking data analysis; The robot dog steering index is used to describe the smoothness of the intelligent robot dog when steering. The specific analysis is as follows: Extract the robot dog’s steering data; The robot dog steering data includes the robot dog steering angle, the robot dog steering speed, the robot dog body length, and the robot dog weight; The robot dog steering angle is used to describe the angle at which the intelligent robot dog turns its head to drive its body to turn when executing the instruction; The robot dog turning index is obtained based on the robot dog turning data analysis; The robot dog motion index is used to describe the working condition of the intelligent robot dog, and the specific analysis is as follows: Extract the robot dog walking index and the robot dog turning index; According to the objective weighting method, the weight ratio of the robot dog walking index to the robot dog movement index and the weight ratio of the robot dog turning index to the robot dog movement index are set; The robot dog movement index is obtained according to the robot dog walking index, the weight ratio of the robot dog walking index to the robot dog movement index, the robot dog turning index, and the weight ratio of the robot dog turning index to the robot dog movement index; The robot dog motion index is processed as follows: Wherein, A represents the robot dog turning index, A' represents the weight ratio of the robot dog turning index to the robot dog motion index, Z represents the robot dog walking index, Z' represents the weight ratio of the robot dog walking index to the robot dog motion index, e represents a natural constant, and D represents the robot dog motion index; The robot dog walking index is processed by the expression: Z=cosh -1 [born -1 (c*v*u)+1] In the formula, c represents the leg length of the robot dog, v represents the walking speed of the robot dog, u represents the friction coefficient of the robot dog's joints, and Z represents the walking index of the robot dog; The robot dog steering index is specifically expressed as: In the formula, g represents the steering angle of the robot dog, v' represents the steering speed of the robot dog, c' represents the length of the robot dog, m represents the weight of the robot dog, and A represents the steering index of the robot dog.
2. The intelligent product quality risk monitoring method based on convolutional neural network as claimed in claim 1, It is characterized in that The specific analysis of the pre-processing based on the appearance image data, the motion performance data, and the robot dog instruction data is as follows: Extract appearance image data, motion performance data, and robot dog command data; The appearance image data is resized to obtain appearance image data of uniform size; De-noising the resized appearance image data, motion performance data, and robot dog command data to eliminate noise in the image; By standardizing and normalizing the denoised appearance image data, motion performance data, and robot dog command data, appearance image data, motion performance data, and robot dog command data of uniform scale and range can be obtained.
3. The intelligent product quality risk monitoring method based on convolutional neural network as claimed in claim 1, It is characterized in that The robot dog appearance damage inspection is used to analyze the appearance damage of the intelligent robot dog, and the specific analysis is as follows: Extracting preprocessed appearance image data; The preprocessed appearance image data is input into a convolutional neural network model based on the appearance image of the intelligent robot dog for training; If the convolutional neural network model based on the appearance image of the intelligent robot dog determines that the appearance of the intelligent robot dog is abnormal, an early warning is issued to notify relevant personnel, and the working status of the robot dog is not checked; If the convolutional neural network model based on the appearance image of the intelligent robot dog determines that the appearance of the intelligent robot dog is normal, the working status of the robot dog is checked.
4. The intelligent product quality risk monitoring method based on convolutional neural network as claimed in claim 1, It is characterized in that The robot dog interaction check is used to analyze the response of the intelligent robot dog to the command. The specific process is as follows: Extract the pre-processed robot dog instruction data; The robot dog command data includes command response time, robot dog movement distance, and robot dog steering angle; The command response time is used to describe the time interval from when the intelligent robot dog receives the command to when it starts to execute the command; The robot dog movement distance is used to describe the distance that the intelligent robot dog moves forward or backward when executing the instruction; The robot dog interaction index is obtained by analyzing the robot dog instruction data.
5. The intelligent product quality risk monitoring method based on convolutional neural network as claimed in claim 4, Features: The robot dog interaction index is used to describe the accuracy of the intelligent robot dog's response to instructions; The specific formula of the robot dog interaction index is: Where t represents the command response time, represents the command response standard value, x represents the robot dog's movement distance, g represents the robot dog's steering angle, and J represents the robot dog's interaction index; Get the robot dog interaction standard value; The robot dog interaction standard value is used to describe the minimum accurate value of the intelligent robot dog's response to the command; If the robot dog interaction index is less than the robot dog interaction standard value, the intelligent robot dog fails to execute the command in a timely and accurate manner, and an early warning is issued to notify relevant personnel; If the robot dog interaction index is not less than the robot dog interaction standard value, the intelligent robot dog executes the instructions promptly and accurately without issuing an early warning.
6. An intelligent product quality risk monitoring system based on convolutional neural network, used to execute the intelligent product quality risk monitoring method based on convolutional neural network according to any one of claims 1 to 5, It is characterized in that include: Data acquisition and preprocessing module, robot dog appearance damage inspection module, robot dog working status inspection module, robot dog interaction inspection module; The data acquisition and preprocessing module is used to acquire the quality risk data of the intelligent robot dog, the quality risk data includes appearance image data, motion performance data, and robot dog instruction data, and preprocess the appearance image data, motion performance data, and robot dog instruction data; The robot dog appearance damage inspection module is used to judge the appearance damage of the intelligent robot dog based on the convolutional neural network model of the intelligent robot dog appearance image, and perform early warning processing according to the appearance damage of the intelligent robot dog; The robot dog working status checking module is used to obtain the robot dog's motion index according to the motion performance data analysis, and to perform early warning processing according to the robot dog's motion index; The robot dog interaction checking module is used to obtain a robot dog interaction index according to the robot dog instruction data analysis, and to perform early warning processing according to the robot dog interaction index.
