Energy supply chain management system based on combination of machine visual perception and energy production and marketing relationship

Abstract

The invention provides an energy supply chain management system based on combination of machine visual perception and an energy production and marketing relationship, and relates to the technical field of energy internet, the system comprises a perception layer, data acquired by the perception layer is uploaded to a platform layer through a network layer, and the platform layer analyzes the data and then issues an analysis result to an application layer. According to the energy management system, the energy management system integrating equipment state real-time monitoring, data quality self-evaluation, multi-source information deep fusion and intelligent productivity prediction is constructed, and the cooperation of front-end optical self-inspection and rear-end intelligent processing is combined, so that the monitoring precision of energy production equipment, the system maintenance efficiency and the scientificity of production and marketing decision are improved.

Description

Technical Field

[0001] The present invention relates to the field of energy internet technology, and in particular to an energy supply chain management system based on the combination of machine vision perception and energy production and sales relationships. Background Art

[0002] When machine vision is applied in energy supply chain management systems, it can use image recognition and analysis technology to conduct real-time monitoring and intelligent diagnosis of equipment status, environmental safety, and operating efficiency in energy production, transportation, consumption, and other links, thereby achieving refined management and optimization of the energy production and sales process. For example, it can prevent malfunctions and downtime by identifying abnormal equipment status, eliminate safety hazards by analyzing the environment around energy facilities, and optimize production capacity scheduling by monitoring energy flow and equipment operating parameters, ultimately improving the reliability, safety, and economy of the energy supply chain.

[0003] However, in the existing technology, existing energy supply chain management systems widely use cameras for visual monitoring. However, cameras installed outdoors are easily affected by meteorological factors such as strong winds and heavy rains, causing jitter and displacement, resulting in distortion of the collected images and reduced data reliability. The use of software algorithms for post-image stabilization cannot evaluate data quality from the source, and the effect is limited under high-precision requirements, so there is room for improvement. Summary of the Invention

[0004] The purpose of the present invention is to address the shortcomings of the existing technology by constructing an energy management system that integrates real-time monitoring of equipment status, self-assessment of data quality, deep integration of multi-source information and intelligent prediction of production capacity, and combining the collaboration of front-end optical self-inspection and back-end intelligent processing to improve the monitoring accuracy of energy production equipment, system maintenance efficiency and the scientific nature of production and marketing decisions.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an energy supply chain management system based on the combination of machine vision perception and energy production and sales relationships, including a perception layer, wherein the data acquired by the perception layer is uploaded to the platform layer via the network layer, and the platform layer analyzes the data and sends the analysis results to the application layer;

[0006] The perception layer includes a visual perception unit for monitoring energy production equipment, a visual unit self-test module for monitoring the visual perception unit, and an IoT sensing unit for energy production data;

[0007] The network layer is used to transmit image data collected by the perception layer, data from the visual unit self-test module, and energy production data to the platform layer;

[0008] The platform layer includes a data pool for storing data, a machine vision analysis module for analyzing image data, a pre-screening module for judging image quality, and a collaborative analysis module for comprehensive analysis of perception layer data.

[0009] As a preferred embodiment, the visual unit self-inspection module includes a laser transmitter installed on the camera and a cylindrical laser receiver installed on the camera mounting structure, the cylindrical laser receiver includes a bottom laser receiver array, a side laser receiver array is provided on the side of the bottom laser receiver array, the axis of the bottom laser receiver is overlapped with the axis of the side laser receiver array, a laser reflecting surface is provided on a side surface of the bottom laser receiver close to the side laser receiver array, the infrared laser emitted by the laser transmitter is refracted and reflected toward the laser reflecting surface, part of the laser reflected by the laser reflecting surface is directed toward the side laser receiver array, and part of the laser refracted by the laser reflecting surface is directed toward the bottom laser receiver array.

[0010] As a preferred embodiment, after the pre-screening module obtains the image data of the visual perception unit and the infrared perception data of the visual unit self-test module, it calculates the laser path according to the point position data of the side laser receiver array and the point position data of the bottom laser receiver array in the visual unit self-test module, thereby calculating the change in the orientation angle of the laser emitter, and calculates the specific offset data of the image data according to the orientation angle change data of the laser emitter. According to the preset offset threshold, the image data is marked in three levels: level one is image data with an offset less than or equal to 3%, level two is image data with an offset greater than 3% and less than or equal to 5%, and level three is image data with an offset greater than 5%.

[0011] As a preferred embodiment, the specific logic of the offset data calculation of the pre-screening module is:

[0012] First, the pre-screening module reads the spot coordinates of the side laser receiver array and the bottom laser receiver array. Then, the pre-screening module calculates the horizontal deflection angle based on the spot coordinate data of the side laser receiver array. The pre-screening module calculates the pitch angle based on the spot coordinate data of the bottom laser receiver array. The calculation result of the horizontal deflection angle is used as the horizontal resolution calculation parameter of the image, and the calculation result of the pitch angle is used as the vertical resolution calculation parameter of the image. The pixel values ​​that are expected to be offset in the horizontal and vertical directions of the center point of the image are calculated respectively, and then the modulus of the overall offset is calculated to obtain the percentage of the offset vector relative to the diagonal size of the entire image.

[0013] As a preferred embodiment, the machine vision analysis module receives image data filtered by the pre-screening module, performs conventional image preprocessing on the first-level marked image according to the offset marking of the image data by the pre-screening module, performs image correction on the second-level marked image and preprocesses the angle correction according to the offset data, performs time axis marking on the third-level marked image and sends it to the data pool for processing, then selects the corresponding comparison image according to the camera number corresponding to the image data, compares and analyzes the image data with the comparison image, and outputs the results to the data pool and the collaborative analysis module.

[0014] As a preferred embodiment, the collaborative analysis module receives the image data analyzed by the machine vision analysis module and the perception data of the Internet of Things sensing unit, then performs data cleaning and spatiotemporal alignment on the analyzed image data and the perception data, and then associates, matches and extracts features of the analyzed image data and the perception data according to the preset logic, and then performs collaborative reasoning and predictive analysis based on the preset rules, and generates a prediction report based on the results of the predictive analysis and sends it to the application layer.

[0015] As a preferred embodiment, the camera of the visual perception unit of the perception unit and the bottom laser receiver array and the side laser receiver array of the visual unit self-inspection module are grouped and numbered when the system is constructed, and each group of the visual perception unit and the visual unit self-inspection module is assigned a number.

[0016] As a preferred embodiment, the visual perception unit and the visual unit self-checking module collect comparative images when the installation is completed, and upload the collected comparative images to a data pool for storage.

[0017] As a preferred implementation, the application layer is used to provide an interactive interface for users.

[0018] Compared with the prior art, the advantages and positive effects of the present invention are:

[0019] 1. The present invention improves the monitoring accuracy of energy production equipment, system maintenance efficiency, and the scientific nature of production and marketing decisions by constructing an energy management system that integrates real-time monitoring of equipment status, self-assessment of data quality, deep integration of multi-source information, and intelligent prediction of production capacity. It combines the collaboration of front-end optical self-inspection and back-end intelligent processing.

[0020] 2. This invention achieves high-precision digital representation of camera offset by constructing a complete mathematical conversion chain from physical light spot displacement to image pixel offset, combining the deep integration of optical measurement and geometric calculation. This not only provides accurate correction parameters for subsequent image processing, but also optimizes the allocation of analysis resources through a hierarchical mechanism, fundamentally improving the reliability of visual monitoring data and the overall efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The module diagram of the energy supply chain management system proposed by the present invention is based on the combination of machine vision perception and energy production and sales relationship;

[0022] Figure 2 The present invention proposes an operational flow chart of an energy supply chain management system based on the combination of machine vision perception and energy production and sales relationships;

[0023] Figure 3 The present invention proposes a schematic diagram of the infrared light path of the visual unit self-inspection module of the energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] Example 1

[0026] like Figure 1-3 As shown, the present invention provides a technical solution: an energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship, including a perception layer, the data acquired by the perception layer is uploaded to the platform layer through the network layer, and the platform layer sends the analysis results to the application layer after analysis. Specifically, the perception layer collects visual information of energy production equipment, camera's own status data and physical energy data in real time, and the network layer serves as a transmission channel to reliably upload these heterogeneous data to the platform layer. The platform layer serves as an intelligent center, and through the storage capacity of the data pool, the image quality assessment and grading capability of the pre-screening module, the image recognition and processing capability of the machine vision analysis module, and the multi-source data fusion and predictive analysis capability of the collaborative analysis module, the data is deeply processed, and finally the analysis results and prediction reports are notified to the user in the form of an interactive interface through the application layer.

[0027] The perception layer includes a visual perception unit for monitoring energy production equipment, a visual unit self-test module for monitoring the visual perception unit, and an Internet of Things sensor unit for energy production data. The camera of the visual perception unit of the perception unit and the bottom laser receiver array and the side laser receiver array of the visual unit self-test module are grouped and numbered when the system is constructed. Each group of the visual perception unit and the visual unit self-test module is assigned a number. When the installation is completed, the visual perception unit and the visual unit self-test module perform comparative image acquisition and upload the acquired comparative image to the data pool for storage. Specifically, through the coordinated deployment of the visual perception unit, the visual unit self-test module and the Internet of Things sensor unit in the perception layer, a three-dimensional data perception system integrating equipment monitoring, self-monitoring and energy data acquisition is constructed, and by integrating the camera of the visual perception unit with the visual unit self-test module The bottom laser receiver array and the side laser receiver array are bound to the same numbered group during system implementation, ensuring that the image data in the subsequent data stream has a unique correlation with the corresponding self-test status data, so that it can be accurately traced during platform layer processing. At the same time, the comparison images collected and uploaded to the data pool immediately after the installation is completed can provide a standard reference system for the machine vision analysis module. By using the laser spot coordinates to calculate the camera offset, continuous monitoring of the self-test module and real-time data supplementation of the Internet of Things sensor unit are achieved. It not only realizes the visual monitoring of the operating status of energy production equipment, but also through the pre-solidified equipment grouping relationship and reference image, the platform layer can efficiently and accurately process each frame of image and each sensor data during subsequent image quality judgment, offset correction, multi-source data fusion and analysis, greatly improving the data consistency and decision reliability of the entire energy supply chain management system.

[0028] The network layer is used to transmit the image data collected by the perception layer, the data of the visual unit self-inspection module and the energy production data to the platform layer. Specifically, the network layer serves as the core data transmission channel connecting the perception layer and the platform layer. It can adopt high-bandwidth, low-latency communication protocols, such as industrial Ethernet or 5G technology, to transmit multi-source heterogeneous data collected by the perception layer, including high-resolution image streams generated by the visual perception unit, laser spot coordinates and equipment status data generated by the visual unit self-inspection module, and real-time energy production data collected by the Internet of Things sensor unit. The network layer can use the built-in data preprocessing and priority scheduling mechanism to implement compression encoding for large-flow image data to reduce bandwidth occupancy, and use high-priority real-time transmission queues for self-inspection status data and energy sensor data to ensure their timeliness and integrity. At the same time, timestamp synchronization technology can be used to ensure that all data maintain temporal and spatial consistency during transmission.

[0029] The platform layer includes a data pool for storing data, a machine vision analysis module for analyzing image data, a pre-screening module for judging image quality, and a collaborative analysis module for comprehensively analyzing perception layer data. The pre-screening module performs three-level marking on image data according to the offset of the image. Specifically, through the coordinated operation of the data pool, the machine vision analysis module, the pre-screening module and the collaborative analysis module, a multi-level processing flow from data storage, quality judgment to intelligent analysis is constructed. The pre-screening module can calculate the horizontal deflection angle and pitch angle of the laser emitter according to the spot coordinate data by receiving the visual perception unit image data uploaded by the network layer and the laser perception data of the visual unit self-test module, and then derive the pixel offset value of the image center point in the horizontal and vertical directions, and obtain the accurate overall offset percentage by calculating the ratio of the modulus of the offset vector to the size of the image diagonal, and then automatically grade and mark the image data according to the preset offset threshold.

[0030] The application layer is used to provide an interactive interface for users. Specifically, the application layer can realize the function of receiving and intuitively presenting various analysis results issued by the platform layer through an interactive interface that integrates data visualization, real-time monitoring, early warning notification and decision support functions. The application layer can use a highly interactive front-end framework based on the Web to build a customizable dashboard, and connect with the data pool and analysis module of the platform layer in real time through the API gateway, dynamically rendering visual charts, real-time video streams and early warning information panels of the operating status of energy equipment, so that users can reversely trigger re-analysis requests or model optimization instructions of the platform layer through interface interaction.

[0031] Furthermore, the machine vision analysis module receives the image data filtered by the pre-screening module, performs conventional image preprocessing on the first-level marked image according to the offset mark of the image data by the pre-screening module, performs image correction on the second-level marked image and performs angle correction preprocessing according to the offset data, performs time axis marking on the third-level marked image and sends it to the data pool for processing, then selects the corresponding comparison image according to the camera number corresponding to the image data, compares and analyzes the image data with the comparison image, and outputs the result to the data pool and the collaborative analysis module. Specifically, the machine vision analysis module receives the image data with the offset level marked by the pre-screening module, and first implements a differentiated processing flow according to different marking levels: for the first-level marked image, conventional image preprocessing is directly performed, including noise filtering, contrast enhancement and edge sharpening. For the second-level marked images, it is necessary to first perform geometric correction and angle correction through the affine transformation algorithm or the perspective transformation algorithm according to the horizontal deflection angle and pitch angle data calculated by the pre-screening module to eliminate the distortion caused by the change of camera posture, and then perform regular preprocessing. For the third-level marked images, the analysis process is suspended, and an accurate timestamp is automatically added and directly archived to the data pool as the original data for abnormal equipment status or later tracing. Subsequently, the machine vision analysis module calls the corresponding numbered reference comparison image collected during system initialization from the data pool according to the camera number associated with the image data, and performs comparison analysis through feature point matching and deep learning models to identify abnormal status, operating parameters or environmental changes of energy production equipment. Finally, the analysis results are synchronously output to the data pool for persistent storage, and pushed to the collaborative analysis module in real time for multi-source data fusion.

[0032] Furthermore, the collaborative analysis module receives the image data analyzed by the machine vision analysis module and the perception data of the Internet of Things sensor unit, then performs data cleaning and spatiotemporal alignment on the analyzed image data and perception data, and then associates, matches and extracts features of the analyzed image data and perception data according to a preset logic, and then performs collaborative reasoning and predictive analysis based on preset rules, and generates a prediction report based on the results of the predictive analysis and sends it to the application layer. Specifically, the collaborative analysis module serves as the data fusion and intelligent decision-making center of the platform layer. By receiving the image data processed by the machine vision analysis module and the real-time perception data of the Internet of Things sensor unit, the collaborative analysis module first performs data cleaning to eliminate outliers and fill missing values ​​for the heterogeneous data, and performs spatiotemporal alignment to ensure the consistency of the visual data and the sensor data in the time and space dimensions through unified timestamp and spatial coordinate mapping. Then, based on the preset association rules, the image features and sensor parameters are matched and fused, and the feature extraction algorithm is used to mine the deep correlation features in the multi-source data. Finally, based on the preset rule library or machine learning model, collaborative reasoning and predictive analysis are performed, and the analysis results are generated into a structured prediction report, which is sent to the application layer through the API interface.

[0033] In this embodiment, the perception layer constructs a multi-source data collection foundation, in which the visual perception unit monitors the equipment operation status, the visual unit self-test module monitors the camera posture changes in real time through the designed optical monitoring mechanism, the Internet of Things sensor unit collects energy production parameters, and establishes a data traceability foundation through the equipment group numbering and reference image collection and storage during system initialization; the platform layer constitutes the intelligent processing core, the data pool centrally stores the original data and processed data, the pre-screening module derives the pixel-level offset and grades it according to the threshold by parsing the laser spot coordinates of the self-test module, the machine vision analysis module performs differentiated processing based on this, and retrieves the reference image by number for feature comparison and analysis, the collaborative analysis module integrates the visual analysis results with the sensor data, and after data cleaning and spatiotemporal alignment, uses multi-source feature fusion and prediction algorithms to generate a prediction report; the application layer presents the analysis results.

[0034] Example 2

[0035] like Figure 1-3As shown, the visual unit self-inspection module includes a laser transmitter installed on the camera and a cylindrical laser receiver installed on the camera mounting structure, the cylindrical laser receiver includes a bottom laser receiver array, the side of the bottom laser receiver array is provided with a side laser receiver array, the axis of the bottom laser receiver is overlapped with the axis of the side laser receiver array, and a laser reflecting surface is provided on a side surface of the bottom laser receiver close to the side laser receiver array. The infrared laser emitted by the laser transmitter is refracted and reflected by the laser reflecting surface, part of the laser reflected by the laser reflecting surface is directed toward the side laser receiver array, and part of the laser refracted by the laser reflecting surface is directed toward the bottom laser receiver array. Specifically, the visual unit self-inspection module realizes accurate self-monitoring of the camera posture through optical structure design, the laser transmitter installed on the camera and the laser receiver fixed on the mounting structure are connected. The spatial geometry measurement system consists of a cylindrical laser receiver on the structure: the infrared laser beam emitted by the laser transmitter is directed to the laser reflecting surface on the bottom laser receiver array, which can be made by semi-transparent and semi-reflective optical coating. The laser reflecting surface reflects and refracts at the same time. The reflected part of the laser is directed to the side laser receiver array at a specific angle, and the refracted part of the laser is directly transmitted to the bottom laser receiver array. The axis of the bottom laser receiver array and the side laser receiver array coincide with each other to ensure the uniformity of the measurement reference. When the camera changes its posture, the laser transmitter moves with the camera, resulting in a change in the incident light path, thereby causing the coordinates of the light spots on the bottom array and the side array to shift. By analyzing the light spot displacement data on the two mutually perpendicular laser receiver arrays, the angular offset of the camera in the three-dimensional space can be calculated, providing a quantitative basis for the image quality grading of the subsequent pre-screening module.

[0036] Furthermore, after the pre-screening module obtains the image data of the visual perception unit and the infrared perception data of the visual unit self-test module, it calculates the laser path according to the point data of the side laser receiver array and the point data of the bottom laser receiver array in the visual unit self-test module, thereby calculating the change in the orientation angle of the laser emitter, and calculates the specific offset data of the image data according to the orientation angle change data of the laser emitter, and performs three-level marking on the image data according to the preset offset threshold: the first level is image data with an offset less than or equal to 3%, the second level is image data with an offset greater than 3% and less than or equal to 5%, and the third level is image data with an offset greater than 5%. Specifically, the pre-screening module constructs an efficient data filtering mechanism according to the impact of camera posture changes on image quality, and quantifies the impact of camera posture changes on image quality. It not only provides a clear hierarchical processing basis for the subsequent machine vision analysis module, but also significantly reduces the processing overhead of invalid data through pre-quality assessment, thereby improving the resource utilization and analysis reliability of the entire system.

[0037] The specific logic for calculating the offset data of the pre-screening module is as follows:

[0038] First, the pre-screening module reads the spot coordinates of the side laser receiver array and the bottom laser receiver array. Then, the pre-screening module calculates the horizontal deflection angle based on the spot coordinate data of the side laser receiver array. The pre-screening module calculates the pitch angle based on the spot coordinate data of the bottom laser receiver array. The calculated result of the horizontal deflection angle is used as the horizontal resolution calculation parameter of the image, and the calculated result of the pitch angle is used as the vertical resolution calculation parameter of the image. The pixel values ​​expected to be offset in the horizontal and vertical directions of the center point of the image are calculated respectively, and then the modulus of the overall offset is calculated to obtain the percentage of the offset vector relative to the diagonal size of the entire image, specifically:

[0039] According to the horizontal displacement Δx of the light spot of the side laser receiver array and the system focal length parameter f, the horizontal deflection angle θ is calculated:

[0040] θ=arctan(Δx / f);

[0041] According to the vertical displacement Δy of the bottom array spot and the focal length f, the pitch angle φ is calculated:

[0042] φ=arctan(Δy / f);

[0043] Convert an angle to a pixel offset:

[0044] Calculate the horizontal pixel offset ΔP x :

[0045] ΔP x =k x ×tanθ×R, where k x is the horizontal scale factor, R is the image resolution;

[0046] Calculate the vertical pixel offset ΔP y :

[0047] ΔP y =k y ×tanφ×R, where k y is the vertical scale factor;

[0048] Calculate the overall deviation modulus |ΔP|:

[0049] |ΔP|=√(ΔP x ²+ΔP y ²);

[0050] The diagonal pixel length of the image is L diag The percentage deviation δ is obtained by comparison:

[0051] δ=(|ΔP| / Ldiag )×100%;

[0052] Deviation grading: Level 1 is δ≤3%; Level 2 is 3%<δ≤5%; Level 3 is δ>5%.

[0053] In this embodiment, precise quantitative monitoring of the camera's posture is achieved through a designed optical-geometric measurement system. The visual unit self-test module utilizes a laser transmitter mounted on the camera and a cylindrical laser receiver fixed to the mounting structure to form the basis for spatial measurement. The infrared laser emitted by the laser transmitter irradiates the semi-transparent, semi-reflective optically coated reflective surface on the bottom laser receiver array. This reflective surface simultaneously produces reflection and refraction optical phenomena, with the reflected beam projected onto the side laser receiver arrays, while the refracted beam reaches the bottom laser receiver array directly. When the camera's posture changes, the moving laser transmitter causes the incident light path to change, resulting in a light spot displacement on both sides of the array. The pre-screening module analyzes the light spot coordinate data and, in combination with the system focal length parameters, calculates the horizontal deflection and pitch angles, which are then converted into image pixel offsets through a projection geometric model. Finally, the overall offset modulus is calculated and compared with the image diagonal pixel length to obtain a percentage offset, which is then marked at three levels based on the δ value.

[0054] Working principle:

[0055] like Figure 1-3 As shown, before construction, the present invention first formulates the association logic and rules of the perception layer data, then installs the visual perception unit at a preset position, and installs the laser transmitter of the visual unit self-inspection module on the camera of the visual perception unit, and installs the bottom laser receiver array and the side laser receiver array on the mounting structure where the camera bracket is located. After the picture is calibrated, the image data taken by the camera is collected as the comparison image data analyzed by the machine vision analysis module.

[0056] When the present invention is in operation, the visual perception unit obtains the monitoring image data of the energy production equipment, and the Internet of Things sensing unit obtains the operation data and environmental data of the energy production equipment. Then, the image data and the infrared data of the visual unit self-inspection module are uploaded to the pre-screening module for image offset analysis and graded marking. The graded and marked image data are then uploaded to the machine vision analysis module for analysis. The analyzed data of the machine vision analysis module and the data obtained by the Internet of Things sensing unit are then uploaded to the collaborative analysis module for collaborative analysis. Finally, the results of the analysis by the collaborative analysis module are sent to the application layer for user use.

[0057] The above description is merely a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any person skilled in the art may utilize the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes for application in other fields. However, any simple modification, equivalent change, and modification of the above embodiments made in accordance with the technical essence of the present invention without departing from the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An energy supply chain management system based on the integration of machine vision perception and energy production and sales relationships, characterized by: It includes a perception layer, which uploads data acquired by the perception layer to the platform layer through the network layer. The platform layer analyzes the data and sends the analysis results to the application layer; The perception layer includes a visual perception unit for monitoring energy production equipment, a visual unit self-test module for monitoring the visual perception unit, and an IoT sensing unit for energy production data; The network layer is used to transmit image data collected by the perception layer, data from the visual unit self-test module, and energy production data to the platform layer; The platform layer includes a data pool for storing data, a machine vision analysis module for analyzing image data, a pre-screening module for judging image quality, and a collaborative analysis module for comprehensive analysis of perception layer data.

2. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 1 is characterized by: The visual unit self-inspection module includes a laser transmitter installed on the camera and a cylindrical laser receiver installed on the camera mounting structure. The cylindrical laser receiver includes a bottom laser receiver array. The side of the bottom laser receiver array is provided with a side laser receiver array. The axis of the bottom laser receiver overlaps with the axis of the side laser receiver array. A laser reflecting surface is provided on a side surface of the bottom laser receiver close to the side laser receiver array. The infrared laser emitted by the laser transmitter is refracted and reflected toward the laser reflecting surface. Part of the laser reflected by the laser reflecting surface is directed toward the side laser receiver array, and part of the laser refracted by the laser reflecting surface is directed toward the bottom laser receiver array.

3. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 2 is characterized by: After the pre-screening module obtains the image data of the visual perception unit and the infrared perception data of the visual unit self-test module, it calculates the laser path based on the point position data of the side laser receiver array and the point position data of the bottom laser receiver array in the visual unit self-test module, thereby calculating the change in the orientation angle of the laser emitter. Based on the orientation angle change data of the laser emitter, the specific offset data of the image data is calculated, and according to the preset offset threshold, the image data is marked in three levels: level one is image data with an offset less than or equal to 3%, level two is image data with an offset greater than 3% and less than or equal to 5%, and level three is image data with an offset greater than 5%.

4. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 3 is characterized by: The specific logic of the offset data calculation of the pre-screening module is: First, the pre-screening module reads the spot coordinates of the side laser receiver array and the bottom laser receiver array. Then, the pre-screening module calculates the horizontal deflection angle based on the spot coordinate data of the side laser receiver array. The pre-screening module calculates the pitch angle based on the spot coordinate data of the bottom laser receiver array. The calculation result of the horizontal deflection angle is used as the horizontal resolution calculation parameter of the image, and the calculation result of the pitch angle is used as the vertical resolution calculation parameter of the image. The pixel values ​​that are expected to be offset in the horizontal and vertical directions of the center point of the image are calculated respectively, and then the modulus of the overall offset is calculated to obtain the percentage of the offset vector relative to the diagonal size of the entire image.

5. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 3 is characterized by: The machine vision analysis module receives the image data filtered by the pre-screening module, and performs conventional image preprocessing on the first-level marked image according to the offset mark of the image data by the pre-screening module, performs image correction on the second-level marked image and preprocesses the angle correction according to the offset data, and performs time axis marking on the third-level marked image and sends it to the data pool for processing. Then, the corresponding comparison image is selected according to the camera number corresponding to the image data, and after comparing and analyzing the image data with the comparison image, the result is output to the data pool and the collaborative analysis module.

6. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 5 is characterized by: The collaborative analysis module receives the image data analyzed by the machine vision analysis module and the perception data of the Internet of Things sensor unit, then performs data cleaning and spatiotemporal alignment on the analyzed image data and perception data, and then associates, matches and extracts features from the analyzed image data and perception data according to the preset logic, and then performs collaborative reasoning and predictive analysis based on the preset rules, and generates a prediction report based on the results of the predictive analysis and sends it to the application layer.

7. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 1 is characterized by: The camera of the visual perception unit of the perception unit and the bottom laser receiver array and the side laser receiver array of the visual unit self-test module are grouped and numbered when the system is constructed, and each group of the visual perception unit and the visual unit self-test module is assigned a number.

8. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 1 is characterized by: When the visual perception unit and the visual unit self-checking module are installed, comparative image acquisition is performed and the acquired comparative images are uploaded to a data pool for storage.

9. The energy supply chain management system based on the combination of machine vision perception and energy production and sales relationship according to claim 1 is characterized by: The application layer is used to provide an interactive interface for users.

Images