Method and system for detecting safety of motorcycle product processing workshop

Abstract

The invention relates to the technical field of motorcycle product processing workshop safety detection, and aims to provide a method and system for effectively detecting the safety of a motorcycle product processing workshop, and the method comprises the steps: encrypting operation state information, transmitting the encrypted operation state information to a preset server, and carrying out the decryption and verification of the encrypted operation state information; performing data cleaning on the verification operation state information, converting the cleaned data into a preset format, performing data integration on the initial data, and performing data reduction on the integrated data; identifying abnormal data in the target data, wherein the training process comprises the steps of extracting sample features of a training sample, reconstructing reconstruction features of the training sample by a reconstruction network, identifying the abnormal data of the training sample, and calculating training loss of the abnormal data and a real label; and determining an abnormal workshop area of the motorcycle processing workshop, identifying an abnormal event of the abnormal workshop area, and analyzing an abnormal factor of the abnormal event. The safety of the motorcycle product processing workshop can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of safety detection of motorcycle product processing workshops, and in particular to a safety detection method and system for motorcycle product processing workshops. Background Art

[0002] With the development of industrialization, the safety issues of various production workshops have attracted more and more attention, especially the motorcycle product processing workshop, which is full of dangers. It is an important link in motorcycle production. Since the production process involves various mechanical equipment, electrical equipment and chemicals, these equipment may malfunction during operation and cause safety accidents. Therefore, the safety of the motorcycle product processing workshop is of great significance to the smooth progress of the entire production process.

[0003] The safety inspection method of motorcycle product processing workshops is usually implemented by drone inspection. However, when using drone inspection for motorcycle product processing workshop safety inspection, it is usually necessary to manually set the drone inspection area and path in advance, which makes the safety inspection of motorcycle product processing workshops less efficient. Therefore, a more efficient safety inspection method for motorcycle product processing workshops is needed. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a safety detection method and system for a motorcycle product processing workshop, which can effectively improve the safety of the motorcycle product processing workshop.

[0005] In a first aspect, the present invention provides a method for detecting safety of a motorcycle product processing workshop, comprising: Deploy a data sensor in a motorcycle processing workshop to be inspected, use the data sensor to collect operating status information of workshop equipment in the motorcycle processing workshop, encrypt the operating status information based on the data sensor and transmit it to a preset server, so as to perform decryption and verification on the encrypted operating status information based on the server to obtain verified operating status information; Performing data cleaning on the verification operation status information to obtain cleaned data, converting the cleaned data into a preset format to obtain initial data, performing data integration on the initial data to obtain integrated data, and performing data reduction on the integrated data to obtain target data; The abnormal data in the target data is identified by using a trained abnormal data identification model, wherein the training process of the trained abnormal data identification model includes: obtaining training samples and their corresponding true labels, extracting sample features of the training samples by using a convolutional network in a pre-constructed abnormal data identification model, and reconstructing the reconstructed features of the training samples by using a reconstruction network in the abnormal data identification model based on the sample features, identifying the abnormal data in the training samples by using a fully connected network in the abnormal data identification model according to the sample features and the reconstructed features, calculating the training loss of the abnormal data and the true label by using a loss function in the abnormal data identification model, and obtaining the trained abnormal data identification model when the training loss is not greater than a preset loss; According to the abnormal data, an abnormal workshop area of ​​the motorcycle processing workshop is determined, abnormal events in the abnormal workshop area are identified, abnormal factors of the abnormal events are analyzed, and according to the abnormal factors, a safety inspection report of the motorcycle processing workshop is created.

[0006] In a possible implementation manner of the first aspect, encrypting the running status information includes: Identify the data type of the operation status information; Based on the data type, construct a pre-shared key for the operation status information; The operation status information is encrypted using a preset encryption algorithm according to the pre-shared key.

[0007] In a possible implementation manner of the first aspect, the decrypting and verifying the encrypted running status information based on the server to obtain the verified running status data includes: querying the encrypted key of the running status information in the server; Inputting the encrypted running status information and the key into a pre-built symmetric decryption algorithm to perform decryption processing of the encrypted running status information to obtain decrypted running status information, and identifying a hash value and a digital signature of the decrypted running status information; The verification running status information is generated according to the hash value and the digital signature.

[0008] In a possible implementation manner of the first aspect, performing data integration on the initial data to obtain integrated data includes: Determine a data warehouse for the initial data; defining a data integration method of the initial data in the data warehouse; Based on the data integration method, the initial data is mapped to the data warehouse to obtain the integrated data.

[0009] In a possible implementation manner of the first aspect, performing data reduction on the integrated data to obtain target data includes: Sampling the integrated data to obtain data samples; Extracting data features from the data sample; According to the data features, data points in the data sample are grouped to obtain grouped integrated data; The grouped integrated data is compressed using a compression algorithm to obtain the target data.

[0010] In a possible implementation manner of the first aspect, extracting sample features of the training samples using a convolutional network in a pre-built abnormal data recognition model includes: Performing sample preprocessing on the training samples using the data input layer in the convolutional network to obtain preprocessed samples; The preprocessed samples are convoluted using a convolutional computing layer in the convolutional network to obtain convolutional samples, wherein the convolutional computing layer includes: ; in, represents the convolution sample, represents the pre-processed sample, represents the weights in the convolutional computation layer, Represents the bias in the convolutional computation layer; The convolution sample is nonlinearly mapped using a nonlinear layer in the convolution network to obtain a nonlinear mapped sample, wherein the nonlinear layer includes: ; in, represents the nonlinear mapping sample, represents the i-th sample in the convolution sample, represents the weight of the nonlinear layer, represents the bias of the nonlinear layer, represents the activation function; Performing sample pooling processing on the nonlinear mapping samples by using the pooling layer in the convolutional network to obtain pooled samples; The pooled samples are integrated using a fully connected layer in the convolutional network to obtain the sample features, wherein the fully connected layer includes: ; in, represents the sample characteristics, represents the pooled sample, represents the weight of the fully connected layer, represents the bias of the fully connected layer.

[0011] In a possible implementation manner of the first aspect, reconstructing the reconstructed features of the training samples using a reconstruction network in the abnormal data recognition model based on the sample features includes: Performing content reconstruction on the sample features to obtain content reconstruction features; Performing structural reconstruction on the content reconstruction feature to obtain a structural reconstruction feature; Extracting detail features from the structural reconstruction features; Based on the detail features, the reconstruction features are calculated using the following formula: ; in, represents the reconstruction feature, represents the Sigmoid activation function, represents the global maximum pooling layer, represents the global average pooling layer, represents a two-layer fully connected layer, Represents the detailed features.

[0012] In a possible implementation manner of the first aspect, identifying abnormal data in the training sample using a fully connected network in the abnormal data identification model according to the sample feature and the reconstruction feature includes: Matching the sample feature and the reconstructed feature to obtain a matching point pair; Based on the matching point pairs, the sample features and the reconstruction features are fused using the following formula to obtain fused features: ; in, represents the fusion feature, Indicates the first feature points, represents the number of feature points in the matching point pair, represents a sequence of matching point pairs, half of which are the sample features and half are the reconstructed features, and Represents the same feature point, represents the feature points of c in multiple sample feature layers and multiple reconstruction feature layers, that is Represents the feature representation of c in different feature layers; The fusion feature is binary-classified using the perceptron model in the fully connected network to obtain a binary classification value, wherein the perceptron model includes: ; in, represents the binary classification value, represents the fusion feature, represents the weights of the perceptron model, represents the bias of the perceptron model, represents a symbolic function; Calculating the abnormal category probability corresponding to the binary classification value using the activation function in the fully connected network; The abnormal class probability is used to identify abnormal data in the training samples.

[0013] In a possible implementation manner of the first aspect, the identifying an abnormal event in the abnormal workshop area includes: Acquire abnormal data of the abnormal workshop area to obtain target abnormal data; Based on the event scale of the target abnormal data, sorting the target abnormal data to obtain an abnormal data sequence; Extracting the first abnormal data and the remaining abnormal data in the abnormal data sequence; Based on the abnormal fluctuation range of the first abnormal data, determining the abnormal event category of the first abnormal data; Based on the remaining abnormal data, an abnormal event of the abnormal workshop area is determined from the abnormal event category.

[0014] In a second aspect, the present invention provides a safety detection device for a motorcycle product processing workshop, the device comprising: An information verification module is used to deploy a data sensor in a motorcycle processing workshop to be inspected, collect the operation status information of the workshop equipment in the motorcycle processing workshop by using the data sensor, encrypt the operation status information based on the data sensor and transmit it to a preset server, so as to perform decryption and verification on the encrypted operation status information based on the server to obtain the verified operation status information; A data reduction module is used to perform data cleaning on the verification operation status information to obtain cleaned data, convert the cleaned data into a preset format to obtain initial data, perform data integration on the initial data to obtain integrated data, and perform data reduction on the integrated data to obtain target data; An abnormality identification module, used for identifying abnormal data in the target data using a trained abnormal data identification model, wherein the training process of the trained abnormal data identification model includes: obtaining training samples and their corresponding true labels, extracting sample features of the training samples using a convolutional network in a pre-constructed abnormal data identification model, and reconstructing the reconstructed features of the training samples using a reconstruction network in the abnormal data identification model based on the sample features, identifying abnormal data in the training samples using a fully connected network in the abnormal data identification model based on the sample features and the reconstructed features, calculating the training loss of the abnormal data and the true label using a loss function in the abnormal data identification model, and obtaining a trained abnormal data identification model when the training loss is not greater than a preset loss; A report creation module is used to determine the abnormal workshop area of ​​the motorcycle processing workshop based on the abnormal data, identify abnormal events in the abnormal workshop area, analyze abnormal factors of the abnormal events, and create a safety inspection report of the motorcycle processing workshop based on the abnormal factors.

[0015] Compared with the prior art, the technical principle and beneficial effects of this solution are: The present invention deploys data sensors in a motorcycle processing workshop to be inspected, and uses the data sensors to collect operating status information of workshop equipment in the motorcycle processing workshop, so as to obtain equipment parameters in the motorcycle processing workshop and ensure the safety detection premise of the subsequent motorcycle processing workshop. Furthermore, the embodiment of the present invention encrypts the operating status information based on the data sensor and transmits it to a preset server to ensure the information security of the operating status information during transmission, and the server performs decryption and verification of the encrypted operating status information, so as to realize the security verification of the operating status information and ensure the integrity of the operating status information. Furthermore, the present invention obtains cleaned data by performing data cleaning on the verified operating status information. In order to filter out useless information in the inspection operation status information, the speed of subsequent data processing is improved. Furthermore, the present invention obtains initial data by converting the cleaned data into a preset format to unify the data format in the cleaned data, and further improves the processing speed of subsequent data. Furthermore, the present invention obtains integrated data by performing data integration on the initial data to improve the quality and accuracy of the data, and the efficiency of data analysis. Furthermore, the present invention obtains target data by performing data reduction on the integrated data to reduce the storage space of the data to improve the efficiency of data processing. Furthermore, the embodiment of the present invention identifies the abnormal data in the target data by using a trained abnormal data recognition model, so as to use a neural network model to quickly identify the abnormal data. Therefore, the motorcycle product processing workshop safety detection method and system proposed in the embodiment of the present invention can effectively improve the safety of the motorcycle product processing workshop. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0018] Figure 1 A schematic flow chart of a method for detecting safety of a motorcycle product processing workshop provided by one embodiment of the present invention; Figure 2 In one embodiment of the present invention Figure 1 A flowchart of one of the steps of a safety testing method for a motorcycle product processing workshop is provided; Figure 3In one embodiment of the present invention Figure 1 A schematic flow chart of another step of a safety testing method for a motorcycle product processing workshop is provided; Figure 4 A schematic diagram of a module of a motorcycle product processing workshop safety detection device provided by an embodiment of the present invention; Figure 5 A schematic diagram of the internal structure of an electronic device for implementing a safety detection method for a motorcycle product processing workshop provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0020] The embodiment of the present application provides a safety detection method for a motorcycle product processing workshop. In the embodiment of the present application, the execution subject of the safety detection method for a motorcycle product processing workshop includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided in the embodiment of the present application. In other words, the safety detection method for a motorcycle product processing workshop can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0021] Reference Figure 1 FIG. 1 is a flow chart of a method for detecting safety of a motorcycle product processing workshop provided by an embodiment of the present invention. In this embodiment, the method for detecting safety of a motorcycle product processing workshop includes steps S1 to S4: S1. Deploy data sensors in a motorcycle processing workshop to be inspected, use the data sensors to collect operating status information of workshop equipment in the motorcycle processing workshop, encrypt the operating status information based on the data sensors and transmit it to a preset server, so as to perform decryption and verification on the encrypted operating status information based on the server to obtain verified operating status information.

[0022] The present invention deploys data sensors in a motorcycle processing workshop to be inspected, and uses the data sensors to collect operating status information of workshop equipment in the motorcycle processing workshop, so as to obtain equipment parameters in the motorcycle processing workshop and ensure the premise of subsequent safety inspection of the motorcycle processing workshop.

[0023] Among them, the data sensor refers to a device installed in a motorcycle processing workshop, which is used to sense, collect, process and transmit the operating status information of various equipment in the workshop. The operating status information includes various parameters such as temperature, humidity, pressure, vibration, current, voltage, etc. After being collected by the data sensor, it can be transmitted to the server in real time for analysis and processing, so as to monitor and control the workshop production process and improve production efficiency and quality.

[0024] Optionally, the data sensor collects the operating status information of the workshop equipment in the motorcycle processing workshop by sensing various physical quantities in the workshop, converting them into analog signals, and amplifying and digitizing them to finally obtain the operating status information. Among them, the physical quantity refers to the quantity used to describe the physical phenomena or object characteristics in the motorcycle processing workshop, including temperature, humidity, pressure, vibration, current, voltage, etc. The analog signal is a continuously changing electrical signal, whose size and shape change with time. The electrical signal refers to the changing signal of current or voltage transmitted in the circuit, which can carry information. The digital processing is the process of converting the analog signal into a digital signal. The digital signal is a discrete, digitized signal, which is composed of a series of discrete digital samples. The advantage of the digital signal is that it can be directly processed and stored by a digital circuit, and has high accuracy and stability. The digital processing process includes three steps: sampling, quantization and encoding. Among them, the sampling is to sample the analog signal at a certain time interval and convert it into a series of values ​​at discrete time points. The quantization is to quantize the sampled signal with a certain accuracy and convert it into a series of discrete values. The encoding is to encode the quantized signal according to a certain encoding method and convert it into the digital signal.

[0025] Furthermore, the embodiment of the present invention encrypts the operating status information based on the data sensor and transmits it to a preset server to ensure the information security of the operating status information during transmission, and the server performs decryption and verification of the encrypted operating status information, so as to achieve security verification of the operating status information and ensure the integrity of the operating status information. The server is a computer system for receiving and processing the sensor data, which can generate analysis reports and provide real-time alarm functions, and the verified operating status information refers to the information after checking and verifying the operating status information.

[0026] Furthermore, in an optional embodiment of the present invention, the transmission of the operating status information can be achieved by creating an Internet of Things gateway between the data sensor and the preset server. The Internet of Things gateway is a bridge connecting the data sensor and the server, which can process and convert the operating status information collected by the data sensor, and then transmit the operating status information to the server for storage and analysis. At the same time, the Internet of Things gateway can also send control instructions to the data sensor to achieve remote control and management of the data sensor.

[0027] As an optional embodiment of the present invention, see Figure 2 As shown, the encrypting the running status information includes: S201, identifying the data type of the operation status information; S202: constructing a pre-shared key for the operation status information based on the data type; S203: Encrypt the running status information according to the pre-shared key and using a preset encryption algorithm.

[0028] Among them, the data type refers to the information identity dimension used to characterize the operating status information, such as medical, financial and other data types. The preset encryption algorithm includes a symmetric encryption algorithm. The symmetric encryption algorithm is an encryption method that uses the same key for encryption and decryption. The pre-shared key is used to establish a symmetric key encryption protocol for secure communication. In the exchange protocol, the communicating parties share a key in advance. When establishing communication, the communicating parties use the key for identity authentication and key exchange to ensure the security of communication.

[0029] Further, as an embodiment of the present invention, the decryption and verification of the encrypted running status information based on the server to obtain the verified running status data includes: querying the key of the encrypted running status information in the server; inputting the encrypted running status information and the key into a pre-built symmetric decryption algorithm to perform decryption processing of the encrypted running status information to obtain the decrypted running status information, and identifying the hash value and digital signature of the decrypted running status information; generating the verified running status information according to the hash value and the digital signature.

[0030] Among them, the symmetric decryption algorithm is a decryption method that uses the same key for encryption and decryption, the hash value is obtained by converting data of arbitrary length into a fixed-length data binary string through a hash algorithm, and the digital signature is obtained by encrypting the hash value of the data using a private key.

[0031] Optionally, the key of the encrypted running status information can be obtained by querying the exchange protocol of the pre-shared key through a query language, and the query language includes the SQL language. The pre-built symmetric decryption algorithm corresponds to the above-mentioned encryption algorithm, which can be compiled through a programming language, such as Java language. The hash value and the digital signature can be obtained by identifying the information field of the decrypted running status information.

[0032] S2. Perform data cleaning on the verification operation status information to obtain cleaned data, convert the cleaned data into a preset format to obtain initial data, perform data integration on the initial data to obtain integrated data, and perform data reduction on the integrated data to obtain target data.

[0033] The present invention performs data cleaning on the verification operation status information to obtain cleaned data, so as to filter out useless information in the verification operation status information and improve the speed of subsequent data processing. The cleaned data refers to the data after processing and screening the verification operation status information to remove useless or erroneous data. Optionally, the data cleaning step generally includes removing duplicate data, processing missing data, processing abnormal data, standardizing data, and screening data. The removal of duplicate data refers to deduplication of data to avoid interference of the duplicate data to subsequent analysis. The processing of missing data refers to processing the missing data by filling default values, interpolation and other methods. The processing of abnormal data refers to processing the abnormal data by deletion, replacement and other methods. The standardized data refers to standardizing the data so that the data has consistency and comparability. The screening data refers to screening the data to remove useless data and retain useful data.

[0034] Furthermore, the present invention converts the cleaned data into a preset format to obtain initial data, so as to unify the data format in the cleaned data and further improve the processing speed of subsequent data. The initial data refers to data that is converted or merged into a description form suitable for data processing. Optionally, converting the cleaned data into a preset format to obtain initial data includes data abstraction processing, data normalization processing, attribute construction processing, data discretization processing, and data dimension reduction processing. Among them, the data abstraction processing refers to hiding the details of the cleaned data and only exposing the necessary interfaces and methods for easy use and maintenance. The data normalization processing is to convert the cleaned data into a certain specification or standard format. The attribute construction processing refers to the operation and processing of data by constructing and processing attributes in a computer program. The attribute construction processing is used to convert the cleaned data into more useful features for classification, clustering and other operations. The data discretization processing refers to the process of converting continuous data into discrete data. The more commonly used discretization processing methods include equal width method, equal frequency method, and clustering method. The equal width method is to divide the data according to a certain interval width, the equal frequency method is to divide the data according to frequency, and the clustering method is to divide the data according to similarity. The data dimensionality reduction processing refers to converting high-dimensional data into low-dimensional data using the kernel principal component analysis method. Among them, the kernel principal component analysis method is a nonlinear dimensionality reduction method that can convert nonlinear data into low-dimensional linear data.

[0035] Furthermore, the present invention integrates the initial data to obtain integrated data, thereby improving the quality and accuracy of the data and the efficiency of data analysis. The integrated data refers to a data set after data from different data sources are integrated into a unified data warehouse.

[0036] As an embodiment of the present invention, the data integration of the initial data to obtain integrated data includes: determining a data warehouse for the initial data; defining a data integration method for the initial data in the data warehouse; and based on the data integration method, mapping the initial data to the data warehouse to obtain the integrated data.

[0037] Among them, the data warehouse is a centralized database system for storing and managing data, the data integration method refers to a method of using real-time data integration tools to synchronize data between data sources in real time, and the mapping refers to matching the fields in the initial data with the fields in the target database.

[0038] Furthermore, the present invention reduces the integrated data to obtain target data, thereby reducing the storage space of the data and improving the efficiency of data processing.

[0039] As an embodiment of the present invention, the integrated data is subjected to data reduction to obtain target data, including: sampling the integrated data to obtain data samples; extracting data features in the data samples; grouping data points in the data samples according to the data features to obtain grouped integrated data; and compressing the grouped integrated data using a compression algorithm to obtain the target data.

[0040] The data sampling refers to the process of extracting a portion of data from a data set for analysis and processing, which can achieve faster processing of large amounts of data and also reduce computing costs and storage space. Optionally, the grouped integrated data can be compressed by Huffman coding, which refers to representing characters with high frequency of occurrence with shorter codes and characters with low frequency of occurrence with longer codes.

[0041] S3. Using a trained abnormal data recognition model to identify abnormal data in the target data, wherein the training process of the trained abnormal data recognition model includes: obtaining training samples and their corresponding true labels, using a convolutional network in a pre-constructed abnormal data recognition model to extract sample features of the training samples, and based on the sample features, using a reconstruction network in the abnormal data recognition model to reconstruct the reconstructed features of the training samples, using a fully connected network in the abnormal data recognition model to identify abnormal data in the training samples based on the sample features and the reconstructed features, using a loss function in the abnormal data recognition model to calculate the training loss of the abnormal data and the true label, and obtaining a trained abnormal data recognition model when the training loss is not greater than a preset loss.

[0042] The embodiment of the present invention identifies abnormal data in the target data by using a trained abnormal data identification model, so as to use a neural network model to quickly identify abnormal data.

[0043] Optionally, the process of identifying abnormal data in the target data by using the trained abnormal data identification model is implemented by a convolutional network, a reconstruction network, and a fully connected network in the trained abnormal data identification model; In one embodiment of the present invention, the method of extracting sample features of the training samples using a convolutional network in a pre-constructed abnormal data recognition model includes: performing sample preprocessing on the training samples using a data input layer in the convolutional network to obtain preprocessed samples; and performing convolution processing on the preprocessed samples using a convolutional calculation layer in the convolutional network to obtain convolutional samples, wherein the convolutional calculation layer includes: ; in, represents the convolution sample, represents the pre-processed sample, represents the weights in the convolutional computation layer, Represents the bias in the convolutional computation layer; The convolution sample is nonlinearly mapped using a nonlinear layer in the convolution network to obtain a nonlinear mapped sample, wherein the nonlinear layer includes: ; in, represents the nonlinear mapping sample, represents the i-th sample in the convolution sample, represents the weight of the nonlinear layer, represents the bias of the nonlinear layer, and f represents the activation function; The nonlinear mapping samples are subjected to sample pooling processing by using the pooling layer in the convolutional network to obtain pooled samples; and the pooled samples are subjected to sample integration by using the fully connected layer in the convolutional network to obtain the sample features, wherein the fully connected layer includes: ; in, represents the sample characteristics, represents the pooled sample, represents the weight of the fully connected layer, represents the bias of the fully connected layer.

[0044] Optionally, the process of performing sample preprocessing on the training samples by using the data input layer in the convolutional network to obtain the preprocessed samples includes de-averaging, normalization, dimensionality reduction, and whitening; In one embodiment of the present invention, based on the sample features, the reconstruction features of the training samples are reconstructed using the reconstruction network in the abnormal data recognition model, including: performing content reconstruction on the sample features to obtain content reconstruction features; performing structural reconstruction on the content reconstruction features to obtain structural reconstruction features; extracting detail features from the structural reconstruction features; and calculating the reconstruction features based on the detail features using the following formula: ; in, represents the reconstruction feature, represents the Sigmoid activation function, represents the global maximum pooling layer, represents the global average pooling layer, represents a two-layer fully connected layer, Represents the detailed features.

[0045] Optionally, the process of performing content reconstruction on the sample features to obtain content reconstruction features is implemented through a CRM network, wherein the CRM network consists of a four-fold upsampling sub-pixel convolution, an activation function and two convolutional layers, which are used to upsample the sample features to the target dimension.

[0046] Optionally, the process of structurally reconstructing the content reconstruction features to obtain the structural reconstruction features is implemented through an SRM network structure that is the same as the above-mentioned CRM network. It should be noted that CRM is mainly used to restore global information and minimize pixel-by-pixel errors, and the purpose of SRM is to utilize structural loss to maintain good structural information based on the CRM results.

[0047] Optionally, the process of extracting detail features from the structural reconstruction features is implemented by a neural network consisting of a fusion of multiple extended convolutional layers of many different sizes and dimensions plus skip connections; In one embodiment of the present invention, the method of identifying abnormal data in the training sample using a fully connected network in the abnormal data identification model according to the sample features and the reconstruction features includes: matching the sample features and the reconstruction features to obtain matching point pairs; based on the matching point pairs, fusing the sample features and the reconstruction features using the following formula to obtain a fused feature: ; in, represents the fusion feature, represents the rth feature point in the matching point pair, represents the number of feature points in the matching point pair, represents a sequence of matching point pairs, half of which are the sample features and the other half are the reconstructed features, and c represents the same feature point. represents the feature points of c in multiple sample feature layers and multiple reconstruction feature layers, that is Represents the feature representation of c in different feature layers; The fusion feature is binary-classified using the perceptron model in the fully connected network to obtain a binary classification value, wherein the perceptron model includes: ; in, represents the binary classification value, represents the fusion feature, represents the weights of the perceptron model, represents the bias of the perceptron model, represents a symbolic function; Calculating the abnormal category probability corresponding to the binary classification value using the activation function in the fully connected network; identifying abnormal data in the training sample using the abnormal category probability; In one embodiment of the invention, the loss function includes: ; in, represents the training loss, Represents training samples The true label of is 1 for positive class and 0 for negative class. It represents the probability that the abnormal data predicted by training sample j is a positive class.

[0048] S4. Determine the abnormal workshop area of ​​the motorcycle processing workshop based on the abnormal data, identify abnormal events in the abnormal workshop area, analyze abnormal factors of the abnormal events, and create a safety inspection report of the motorcycle processing workshop based on the abnormal factors.

[0049] In one embodiment of the present invention, determining the abnormal workshop area of ​​the motorcycle processing workshop based on the abnormal data includes: identifying the source area and the destination area of ​​the abnormal data from the motorcycle processing workshop; and using the source area and the destination area as the abnormal workshop area.

[0050] In one embodiment of the present invention, the identifying of abnormal events in the abnormal workshop area includes: acquiring abnormal data of the abnormal workshop area to obtain target abnormal data; sorting the target abnormal data based on the event scale of the target abnormal data to obtain an abnormal data sequence; extracting the first abnormal data and the remaining abnormal data in the abnormal data sequence; based on the abnormal fluctuation range of the first abnormal data, determining the abnormal event category of the first abnormal data; based on the remaining abnormal data, determining the abnormal events in the abnormal workshop area from the abnormal event categories.

[0051] Among them, the abnormal fluctuation range refers to the range of the data value, and each range has a corresponding abnormal event category. For example, in the range of 0 to 1000, it is divided into 0 to 500 and 500 to 1000. 0 to 500 corresponds to abnormal event category A, and 500 to 1000 corresponds to abnormal event category B.

[0052] Optionally, the process of determining the abnormal events of the abnormal workshop area from the abnormal event categories based on the remaining abnormal data is: based on the abnormal fluctuation range of the remaining abnormal data, judging the abnormal event category of the remaining abnormal data, and selecting the abnormal event from the intersection of the abnormal event category of the remaining abnormal data and the abnormal event category of the first abnormal data.

[0053] In the embodiment of the present invention, the abnormal factor refers to an abnormal factor or component.

[0054] In one embodiment of the present invention, referring to Figure 3 As shown, the analysis of the abnormal factors of the abnormal event includes: S301, extracting abnormal keywords from the abnormal event; S302, performing keyword clustering on the abnormal keywords to obtain keyword clusters; S303, identifying a keyword cluster in which the number of keywords in the keyword cluster is greater than a preset number, and obtaining a target cluster; S304: Use the keyword category corresponding to the target cluster as the abnormal factor.

[0055] In the embodiment of the present invention, the safety detection report refers to a report composed of abnormal factors.

[0056] It can be seen that the present invention deploys data sensors in the motorcycle processing workshop to be inspected, and uses the data sensors to collect the operating status information of the workshop equipment in the motorcycle processing workshop, so as to obtain the equipment parameters in the motorcycle processing workshop and ensure the safety detection premise of the subsequent motorcycle processing workshop. Furthermore, the embodiment of the present invention encrypts the operating status information based on the data sensor and transmits it to a preset server to ensure the information security of the operating status information during the transmission process, and performs decryption and verification of the encrypted operating status information by the server, so as to realize the security verification of the operating status information and ensure the integrity of the operating status information. Furthermore, the present invention performs data cleaning on the verified operating status information to obtain the cleaned information. Data is collected to filter out useless information in the inspection operation status information, thereby improving the speed of subsequent data processing. Furthermore, the present invention converts the cleaned data into a preset format to obtain initial data, thereby unifying the data format in the cleaned data, thereby further improving the processing speed of subsequent data. Furthermore, the present invention integrates the initial data to obtain integrated data, thereby improving the quality and accuracy of the data and the efficiency of data analysis. Furthermore, the present invention reduces the integrated data to obtain target data, thereby reducing the storage space of the data and improving the efficiency of data processing. Furthermore, an embodiment of the present invention identifies abnormal data in the target data by using a trained abnormal data recognition model, thereby using a neural network model to quickly identify abnormal data. Therefore, a motorcycle product processing workshop safety detection method proposed in an embodiment of the present invention can effectively improve the safety of a motorcycle product processing workshop.

[0057] like Figure 4The figure shows a functional module diagram of the motorcycle product processing workshop safety detection device of the present invention.

[0058] The motorcycle product processing workshop safety detection device 400 of the present invention can be installed in an electronic device. According to the functions to be implemented, the motorcycle product processing workshop safety detection device can include an information verification module 401, a data reduction module 402, an abnormality identification module 403 and a report creation module 404. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0059] In the embodiment of the present invention, the functions of each module / unit are as follows: The information verification module 401 is used to deploy data sensors in the motorcycle processing workshop to be inspected, collect the operating status information of the workshop equipment in the motorcycle processing workshop by using the data sensors, encrypt the operating status information based on the data sensors and transmit it to a preset server, so as to perform decryption and verification on the encrypted operating status information based on the server to obtain the verified operating status information; The data reduction module 402 is used to perform data cleaning on the verification operation status information to obtain cleaned data, convert the cleaned data into a preset format to obtain initial data, perform data integration on the initial data to obtain integrated data, and perform data reduction on the integrated data to obtain target data; The abnormality identification module 403 is used to identify abnormal data in the target data using a trained abnormal data identification model, wherein the training process of the trained abnormal data identification model includes: obtaining training samples and their corresponding true labels, extracting sample features of the training samples using a convolutional network in a pre-constructed abnormal data identification model, and reconstructing the reconstructed features of the training samples using a reconstruction network in the abnormal data identification model based on the sample features, identifying abnormal data in the training samples using a fully connected network in the abnormal data identification model based on the sample features and the reconstructed features, calculating the training loss of the abnormal data and the true label using a loss function in the abnormal data identification model, and obtaining a trained abnormal data identification model when the training loss is not greater than a preset loss; The report creation module 404 is used to determine the abnormal workshop area of ​​the motorcycle processing workshop based on the abnormal data, identify abnormal events in the abnormal workshop area, analyze abnormal factors of the abnormal events, and create a safety inspection report of the motorcycle processing workshop based on the abnormal factors.

[0060] In detail, the modules in the motorcycle product processing workshop safety detection device 400 in the embodiment of the present invention are used in the same manner as above. Figures 1 to 3 The same technical means are used as the motorcycle product processing workshop safety testing method described in, and can produce the same technical effects, so they will not be repeated here.

[0061] like Figure 5 FIG. 1 is a schematic diagram of the structure of an electronic device for implementing a safety detection method for a motorcycle product processing workshop according to the present invention.

[0062] The electronic device may include a processor 50, a memory 51, a communication bus 52 and a communication interface 53, and may also include a computer program stored in the memory 51 and executable on the processor 50, such as a motorcycle product processing workshop safety detection program.

[0063] In some embodiments, the processor 50 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The processor 50 is the control core (ControlUnit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing programs or modules stored in the memory 51 (for example, executing a motorcycle product processing workshop safety detection program, etc.), and calling data stored in the memory 51.

[0064] The memory 51 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 51 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 51 may also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 51 may also include both an internal storage unit of the electronic device and an external storage device. The memory 51 may not only be used to store application software and various types of data installed in the electronic device, such as the code of the database configuration connection program, but also be used to temporarily store data that has been output or is to be output.

[0065] The communication bus 52 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 51 and at least one processor 50, etc.

[0066] The communication interface 53 is used for communication between the above-mentioned electronic device 5 and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.

[0067] Figure 5 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 5 The structure shown does not constitute a limitation on the electronic device, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0068] For example, although not shown, the electronic device may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 50 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, and power status indicators. The electronic device may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0069] It should be understood that the embodiment is for illustrative purposes only and the scope of the patent invention is not limited to this structure.

[0070] The database configuration connection program stored in the memory 51 in the electronic device is a combination of multiple computer programs. When running in the processor 50, it can achieve: Deploy a data sensor in a motorcycle processing workshop to be inspected, use the data sensor to collect operating status information of workshop equipment in the motorcycle processing workshop, encrypt the operating status information based on the data sensor and transmit it to a preset server, so as to perform decryption and verification on the encrypted operating status information based on the server to obtain verified operating status information; Performing data cleaning on the verification operation status information to obtain cleaned data, converting the cleaned data into a preset format to obtain initial data, performing data integration on the initial data to obtain integrated data, and performing data reduction on the integrated data to obtain target data; The abnormal data in the target data is identified by using a trained abnormal data identification model, wherein the training process of the trained abnormal data identification model includes: obtaining training samples and their corresponding true labels, extracting sample features of the training samples by using a convolutional network in a pre-constructed abnormal data identification model, and reconstructing the reconstructed features of the training samples by using a reconstruction network in the abnormal data identification model based on the sample features, identifying the abnormal data in the training samples by using a fully connected network in the abnormal data identification model according to the sample features and the reconstructed features, calculating the training loss of the abnormal data and the true label by using a loss function in the abnormal data identification model, and obtaining the trained abnormal data identification model when the training loss is not greater than a preset loss; According to the abnormal data, an abnormal workshop area of ​​the motorcycle processing workshop is determined, abnormal events in the abnormal workshop area are identified, abnormal factors of the abnormal events are analyzed, and according to the abnormal factors, a safety inspection report of the motorcycle processing workshop is created.

[0071] Specifically, the specific implementation method of the processor 50 for the above computer program can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0072] Furthermore, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium. The storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0073] The present invention further provides a storage medium, wherein the readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement: Deploy a data sensor in a motorcycle processing workshop to be inspected, use the data sensor to collect operating status information of workshop equipment in the motorcycle processing workshop, encrypt the operating status information based on the data sensor and transmit it to a preset server, so as to perform decryption and verification on the encrypted operating status information based on the server to obtain verified operating status information; Performing data cleaning on the verification operation status information to obtain cleaned data, converting the cleaned data into a preset format to obtain initial data, performing data integration on the initial data to obtain integrated data, and performing data reduction on the integrated data to obtain target data; The abnormal data in the target data is identified by using a trained abnormal data identification model, wherein the training process of the trained abnormal data identification model includes: obtaining training samples and their corresponding true labels, extracting sample features of the training samples by using a convolutional network in a pre-constructed abnormal data identification model, and reconstructing the reconstructed features of the training samples by using a reconstruction network in the abnormal data identification model based on the sample features, identifying the abnormal data in the training samples by using a fully connected network in the abnormal data identification model according to the sample features and the reconstructed features, calculating the training loss of the abnormal data and the true label by using a loss function in the abnormal data identification model, and obtaining the trained abnormal data identification model when the training loss is not greater than a preset loss; According to the abnormal data, an abnormal workshop area of ​​the motorcycle processing workshop is determined, abnormal events in the abnormal workshop area are identified, abnormal factors of the abnormal events are analyzed, and according to the abnormal factors, a safety inspection report of the motorcycle processing workshop is created.

[0074] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0075] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0077] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0078] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0079] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0080] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A motorcycle product processing workshop safety detection method and system, characterized in that: The method comprises: Deploy a data sensor in a motorcycle processing workshop to be inspected, use the data sensor to collect operating status information of workshop equipment in the motorcycle processing workshop, encrypt the operating status information based on the data sensor and transmit it to a preset server, so as to perform decryption and verification on the encrypted operating status information based on the server to obtain verified operating status information; Performing data cleaning on the verification operation status information to obtain cleaned data, converting the cleaned data into a preset format to obtain initial data, performing data integration on the initial data to obtain integrated data, and performing data reduction on the integrated data to obtain target data; The abnormal data in the target data is identified by using a trained abnormal data identification model, wherein the training process of the trained abnormal data identification model includes: obtaining training samples and their corresponding true labels, extracting sample features of the training samples by using a convolutional network in a pre-constructed abnormal data identification model, and reconstructing the reconstructed features of the training samples by using a reconstruction network in the abnormal data identification model based on the sample features, identifying the abnormal data in the training samples by using a fully connected network in the abnormal data identification model according to the sample features and the reconstructed features, calculating the training loss of the abnormal data and the true label by using a loss function in the abnormal data identification model, and obtaining the trained abnormal data identification model when the training loss is not greater than a preset loss; According to the abnormal data, an abnormal workshop area of ​​the motorcycle processing workshop is determined, abnormal events in the abnormal workshop area are identified, abnormal factors of the abnormal events are analyzed, and according to the abnormal factors, a safety inspection report of the motorcycle processing workshop is created.

2. The method according to claim 1, characterized in that The encrypting the running status information includes: Identify the data type of the operation status information; Based on the data type, construct a pre-shared key for the operation status information; The operation status information is encrypted using a preset encryption algorithm according to the pre-shared key.

3. The method according to claim 1, characterized in that The decrypting and verifying the encrypted running status information based on the server to obtain the verified running status data includes: querying the encrypted key of the running status information in the server; Inputting the encrypted running status information and the key into a pre-built symmetric decryption algorithm to perform decryption processing of the encrypted running status information to obtain decrypted running status information, and identifying a hash value and a digital signature of the decrypted running status information; The verification running status information is generated according to the hash value and the digital signature.

4. The method according to claim 1, characterized in that: The step of integrating the initial data to obtain integrated data includes: Determine a data warehouse for the initial data; defining a data integration method of the initial data in the data warehouse; Based on the data integration method, the initial data is mapped to the data warehouse to obtain the integrated data.

5. The method according to claim 1, characterized in that The step of performing data reduction on the integrated data to obtain target data includes: Sampling the integrated data to obtain data samples; Extracting data features from the data sample; According to the data features, data points in the data sample are grouped to obtain grouped integrated data; The grouped integrated data is compressed using a compression algorithm to obtain the target data.

6. The method according to claim 1, characterized in that The extracting the sample features of the training samples by using the convolutional network in the pre-built abnormal data recognition model includes: Performing sample preprocessing on the training samples using the data input layer in the convolutional network to obtain preprocessed samples; The preprocessed samples are convoluted using a convolutional computing layer in the convolutional network to obtain convolutional samples, wherein the convolutional computing layer includes: ； in, represents the convolution sample, represents the pre-processed sample, represents the weights in the convolutional computation layer, Represents the bias in the convolutional computation layer; The convolution sample is nonlinearly mapped using a nonlinear layer in the convolution network to obtain a nonlinear mapped sample, wherein the nonlinear layer includes: ； in, represents the nonlinear mapping sample, represents the i-th sample in the convolution sample, represents the weight of the nonlinear layer, represents the bias of the nonlinear layer, and f represents the activation function; Performing sample pooling processing on the nonlinear mapping samples by using the pooling layer in the convolutional network to obtain pooled samples; The pooled samples are integrated using a fully connected layer in the convolutional network to obtain the sample features, wherein the fully connected layer includes: ； in, represents the sample characteristics, represents the pooled sample, represents the weight of the fully connected layer, represents the bias of the fully connected layer.

7. The method according to claim 1, characterized in that The reconstructing the reconstructed features of the training samples using the reconstruction network in the abnormal data recognition model based on the sample features includes: Performing content reconstruction on the sample features to obtain content reconstruction features; Performing structural reconstruction on the content reconstruction feature to obtain a structural reconstruction feature; Extracting detail features from the structural reconstruction features; Based on the detail features, the reconstruction features are calculated using the following formula: ； in, represents the reconstruction feature, represents the Sigmoid activation function, represents the global maximum pooling layer, represents the global average pooling layer, represents a two-layer fully connected layer, Represents the detailed features.

8. The method according to claim 1, characterized in that The step of identifying abnormal data in the training sample using a fully connected network in the abnormal data identification model according to the sample features and the reconstruction features includes: Matching the sample feature and the reconstructed feature to obtain a matching point pair; Based on the matching point pairs, the sample features and the reconstruction features are fused using the following formula to obtain fused features: ； in, represents the fusion feature, represents the rth feature point in the matching point pair, represents the number of feature points in the matching point pair, represents a sequence of matching point pairs, half of which are the sample features and the other half are the reconstructed features, and c represents the same feature point. represents the feature points of c in multiple sample feature layers and multiple reconstruction feature layers, that is Represents the feature representation of c in different feature layers; The fusion feature is binary-classified using the perceptron model in the fully connected network to obtain a binary classification value, wherein the perceptron model includes: ； in, represents the binary classification value, represents the fusion feature, represents the weights of the perceptron model, represents the bias of the perceptron model, represents a symbolic function; and calculating the abnormal category probability corresponding to the binary classification value using the activation function in the fully connected network; The abnormal class probability is used to identify abnormal data in the training samples.

9. The method according to claim 1, characterized in that: The identifying the abnormal event in the abnormal workshop area includes: Acquire abnormal data of the abnormal workshop area to obtain target abnormal data; Based on the event scale of the target abnormal data, sorting the target abnormal data to obtain an abnormal data sequence; Extracting the first abnormal data and the remaining abnormal data in the abnormal data sequence; Based on the abnormal fluctuation range of the first abnormal data, determining the abnormal event category of the first abnormal data; Based on the remaining abnormal data, an abnormal event of the abnormal workshop area is determined from the abnormal event category.

10. A safety detection device for a motorcycle product processing workshop, characterized in that: The device comprises: An information verification module is used to deploy a data sensor in a motorcycle processing workshop to be inspected, collect the operation status information of the workshop equipment in the motorcycle processing workshop by using the data sensor, encrypt the operation status information based on the data sensor and transmit it to a preset server, so as to perform decryption and verification on the encrypted operation status information based on the server to obtain the verified operation status information; A data reduction module is used to perform data cleaning on the verification operation status information to obtain cleaned data, convert the cleaned data into a preset format to obtain initial data, perform data integration on the initial data to obtain integrated data, and perform data reduction on the integrated data to obtain target data; An abnormality identification module, used for identifying abnormal data in the target data using a trained abnormal data identification model, wherein the training process of the trained abnormal data identification model includes: obtaining training samples and their corresponding true labels, extracting sample features of the training samples using a convolutional network in a pre-constructed abnormal data identification model, and reconstructing the reconstructed features of the training samples using a reconstruction network in the abnormal data identification model based on the sample features, identifying abnormal data in the training samples using a fully connected network in the abnormal data identification model based on the sample features and the reconstructed features, calculating the training loss of the abnormal data and the true label using a loss function in the abnormal data identification model, and obtaining a trained abnormal data identification model when the training loss is not greater than a preset loss; A report creation module is used to determine the abnormal workshop area of ​​the motorcycle processing workshop based on the abnormal data, identify abnormal events in the abnormal workshop area, analyze abnormal factors of the abnormal events, and create a safety inspection report of the motorcycle processing workshop based on the abnormal factors.

Images