A visualized peanut processing whole-process monitoring method and system

Abstract

The application provides a peanut processing visualized whole-process monitoring method and system, and belongs to the technical field of supervisory control and data acquisition. The method can accurately identify the overload or underload imbalance state of each process area by comparing the deviation between the personnel comprehensive density value and the equipment comprehensive load value, and completes the accurate screening and matching of cross-area personnel allocation based on the imbalance state, realizes the dynamic and optimized allocation of workshop personnel resources, and realizes the dynamic balance of personnel allocation and equipment load. At the same time, based on the personnel distribution and equipment load data update after personnel scheduling, a dynamic visualized correlation diagram between personnel distribution and equipment load is generated, the dynamic corresponding relationship between the personnel distribution state change and the equipment load change of each process area is intuitively presented on the monitoring interface, so that the management personnel can master the whole-process state of peanut processing in real time and clearly.

Description

Technical Field

[0001] This application relates to the fields of supervision and control and data acquisition technology, specifically to a visual full-process monitoring method and system for peanut processing. Background Technology

[0002] Currently, the peanut processing industry, as a crucial component of deep agricultural product processing, directly impacts food safety, product quality stability, and enterprise production efficiency. In large-scale continuous production environments, ensuring the entire processing flow—from raw material intake to finished product delivery—remains clearly visible and controllable has become a key requirement for guaranteeing product quality and enhancing competitiveness. However, most peanut processing enterprises still rely on manual inspections, scattered equipment and instruments, and post-production records to manage the production process. This approach makes it difficult to accurately grasp the real-time correlation between personnel distribution, equipment load, and raw material flow within the workshop. During peak production periods, a sudden surge in operators at one stage, coupled with staff shortages at adjacent stages, leads to frequent instances of equipment idling or overloading, which are difficult to detect and address on-site immediately. Managers often only gain a general understanding of the problem through post-production statistical reports, disrupting production rhythms, increasing batch-to-batch quality fluctuations, and ultimately reducing peanut processing efficiency.

[0003] The information provided in the background section of this application is only for enhancing the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] In view of this, this application provides a visual full-process monitoring method and system for peanut processing, which can monitor the entire peanut processing process in real time.

[0005] In a first aspect, embodiments of this application provide a visualized full-process monitoring method for peanut processing. The method includes: acquiring multi-angle image data of personnel in each process area of ​​a peanut processing workshop; determining the real-time location coordinates and dwell time of personnel in each process area based on the multi-angle image data; determining the personnel distribution in each process area based on the real-time location coordinates and dwell time, wherein the personnel distribution includes personnel gathering frequency and inter-personal cooperation distance, and determining the comprehensive density value of personnel in each process area based on the personnel gathering frequency and inter-personal cooperation distance; acquiring the operating parameters of equipment in each process area, and determining the load index of equipment in each process area based on the operating parameters, wherein the load index includes real-time processing cycle time. The deviation value, spindle load rate fluctuation amplitude, and peak intensity of vibration spectrum characteristics are used to determine the comprehensive load value of equipment in each process area based on the load index. If the deviation between the comprehensive density value and the comprehensive load value exceeds a preset deviation threshold, the current process area is marked as an overload area. If the deviation between the comprehensive density value and the comprehensive load value does not exceed the preset deviation threshold, the current process area is marked as an underload area, and a list of personnel to be transferred from the overload area to the underload area is determined. The personnel distribution and the load index are updated based on the personnel transfer list, and a dynamic visualization correlation diagram between personnel distribution and equipment load is determined based on the updated personnel distribution and the load index.

[0006] Secondly, embodiments of this application provide a visualized full-process monitoring system for peanut processing. This system includes: a first acquisition module, a first determination module, a second determination module, a second acquisition module, a third determination module, a fourth determination module, and a fifth determination module. The first acquisition module is used to acquire multi-angle image data of personnel in each process area of ​​the peanut processing workshop; the first determination module is used to determine the real-time location coordinates and dwell time of personnel in each process area based on the multi-angle image data; the second determination module is used to determine the personnel distribution in each process area based on the real-time location coordinates and dwell time, the personnel distribution including personnel gathering frequency and inter-personal cooperation distance, and to determine the comprehensive density value of personnel in each process area based on the personnel gathering frequency and inter-personal cooperation distance; the second acquisition module is used to acquire the operating parameters of equipment in each process area, and to determine the load index of equipment in each process area based on the operating parameters, the load index including the real-time deviation value of processing cycle time and the fluctuation range of spindle load rate. The system includes a third determining module, which determines the comprehensive load value of equipment in each process area based on the load index; a fourth determining module, which marks the current process area as an overload area if the deviation between the comprehensive density value and the comprehensive load value exceeds a preset deviation threshold, and marks the current process area as an underload area if the deviation between the comprehensive density value and the comprehensive load value exceeds the preset deviation threshold, and determines a list of personnel to be transferred from the overload area to the underload area; and a fifth determining module, which updates the personnel distribution and the load index based on the list of personnel to be transferred, and determines a dynamic visualization correlation diagram between personnel distribution and equipment load based on the updated personnel distribution and the load index.

[0007] This application provides a visualized full-process monitoring method and system for peanut processing. The method determines the comprehensive personnel density value by analyzing multi-angle image data of personnel in various process areas of the peanut processing workshop, and integrates multi-dimensional operating parameters of equipment in each process area into a comprehensive equipment load value. Based on this, by comparing the deviation between the comprehensive personnel density value and the comprehensive equipment load value, it can accurately identify the overload or underload imbalance state of each process area, and accurately screen and match personnel transferred across areas based on the imbalance state. This achieves dynamic and optimized allocation of workshop personnel resources, effectively solving the problems of relying on experience-based judgment for personnel scheduling in traditional production processes, which easily leads to raw material backlog or equipment waiting for operation. It achieves a dynamic balance between personnel allocation and equipment load, significantly improving the efficiency of human-machine collaborative operation, and fundamentally avoiding disruption of the entire production line's production rhythm due to local process load imbalances. Meanwhile, based on the updated personnel distribution and equipment load data after personnel scheduling, a dynamic and visual correlation diagram between personnel distribution and equipment load is generated. This intuitively presents the dynamic correspondence between changes in personnel distribution status and changes in equipment load in each process area on the monitoring interface, allowing managers to grasp the production status of the entire process in real time and clearly. This solves the information gap problem in the traditional monitoring mode, realizes clear visibility, controllability and adjustability of the entire peanut processing process, effectively reduces product quality fluctuations between batches, ensures production stability, and significantly improves the overall production efficiency and market competitiveness of the enterprise. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating a visualized full-process monitoring method for peanut processing provided in an exemplary embodiment of this application.

[0010] Figure 2 This is a flowchart illustrating a method for visualizing the entire process of peanut processing, provided in another exemplary embodiment of this application.

[0011] Figure 3 This is a flowchart illustrating a method for visualizing the entire process of peanut processing, provided in another exemplary embodiment of this application.

[0012] Figure 4 This is a flowchart illustrating a method for visualizing the entire process of peanut processing, provided in another exemplary embodiment of this application.

[0013] Figure 5 This is a flowchart illustrating a method for visualizing the entire process of peanut processing, provided in another exemplary embodiment of this application.

[0014] Figure 6 This is a flowchart illustrating a method for visualizing the entire process of peanut processing, provided in another exemplary embodiment of this application.

[0015] Figure 7 This is a flowchart illustrating a method for visualizing the entire process of peanut processing, provided in another exemplary embodiment of this application.

[0016] Figure 8 This is a flowchart illustrating a method for visualizing the entire process of peanut processing, provided in another exemplary embodiment of this application. Detailed Implementation

[0017] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this application will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this application.

[0018] The terms “a,” “one,” and “the” are used to indicate the existence of one or more elements / components / etc.; the terms “including” and “having” are used to indicate an open-ended inclusion and that other elements / components / etc. may exist in addition to those listed. The terms “first” and “second” are used only as markers and are not a limitation on the number of objects.

[0019] Currently, the peanut processing industry, as a crucial component of deep agricultural product processing, directly impacts food safety, product quality stability, and enterprise production efficiency. In large-scale continuous production environments, ensuring the entire processing flow—from raw material intake to finished product delivery—remains clearly visible and controllable has become a key requirement for guaranteeing product quality and enhancing competitiveness. However, most peanut processing enterprises still rely on manual inspections, scattered equipment and instruments, and post-production records to manage the production process. This approach makes it difficult to accurately grasp the real-time correlation between personnel distribution, equipment load, and raw material flow within the workshop. During peak production periods, a sudden surge in operators at one stage, coupled with staff shortages at adjacent stages, leads to frequent instances of equipment idling or overloading, which are difficult to detect and address on-site immediately. Managers often only gain a general understanding of the problem through post-production statistical reports, disrupting production rhythms, increasing batch-to-batch quality fluctuations, and ultimately reducing peanut processing efficiency.

[0020] Specifically, peanut processing involves multiple steps, including washing, grading, roasting, cooling, screening, and packaging. Each step has significantly different processing speeds and personnel requirements, and these differences change in real time depending on factors such as raw material moisture content, kernel size, and impurity content. Personnel movement within the workshop is highly dynamic, requiring operators to frequently move between different workstations to balance workloads. However, existing monitoring methods cannot effectively link the actual location and dwell time of personnel with the real-time operating status of equipment in each process. This means that while managers can see whether the equipment is operating, they cannot see "who is in which position, what they are doing, or whether it matches the current production rhythm," thus failing to promptly detect imbalances between personnel allocation and production capacity demands.

[0021] For example, during peak hours, when a particular process requires more personnel due to a concentration of incoming materials, without a clear indication of the match between the personnel gathering area and the equipment's processing capacity, managers will find it difficult to quickly determine whether personnel need to be redeployed from other processes or adjusted the raw material delivery schedule. This information gap leads to low efficiency in human-machine collaboration, easily causing backlogs and waiting times in some processes, which in turn affects the smooth operation of the entire production line and the consistency of product quality.

[0022] Therefore, how to achieve dynamic visualization and correlation between real-time personnel distribution and equipment load in peanut processing workshops based on the technical architecture of monitoring and control and data acquisition systems, and accurately capture the matching relationship between the two, has become a technical problem that needs to be solved to improve the monitoring capability of the entire peanut processing process and ensure the stability of peanut production and product quality.

[0023] This application provides a method for visually monitoring the entire peanut processing process, such as... Figure 1The method described here is a visualized, end-to-end monitoring method for peanut processing. This method may include the following steps: Step S110: Obtain multi-angle image data of personnel in each process area of ​​the peanut processing workshop; Step S120: Determine the real-time location coordinates and dwell time of personnel in each process area based on multi-angle image data; Step S130: Determine the personnel distribution in each process area based on real-time location coordinates and dwell time. The personnel distribution includes personnel gathering frequency and inter-personal collaboration distance. Determine the comprehensive personnel density value in each process area based on personnel gathering frequency and inter-personal collaboration distance. Step S140: Obtain the operating parameters of the equipment in each process area, and determine the load index of the equipment in each process area based on the operating parameters. The load index includes the real-time deviation of the processing cycle time, the fluctuation amplitude of the spindle load rate, and the peak intensity of the vibration spectrum characteristics. Step S150: Determine the comprehensive load value of equipment in each process area based on the load index; Step S160: If the deviation between the comprehensive density value and the comprehensive load value exceeds the preset deviation threshold, the current process area is marked as an overloaded area; if the deviation between the comprehensive density value and the comprehensive load value exceeds the preset deviation threshold, the current process area is marked as an underloaded area, and a list of personnel to be transferred from the overloaded area to the underloaded area is determined. Step S170: Update the personnel distribution and load indicators according to the list of transferred personnel, and determine the dynamic visualization relationship between personnel distribution and equipment load based on the updated personnel distribution and load indicators.

[0024] According to the peanut processing visualization monitoring method provided in this application, this method relies on a supervision and control and data acquisition system to achieve end-to-end data acquisition, data analysis, data scheduling, and data visualization. It can accurately identify the overload or underload imbalance in each process area by comparing the deviation between the comprehensive personnel density value and the comprehensive equipment load value. Based on the imbalance, it can accurately screen and match personnel transferred across areas, realizing dynamic and optimized allocation of workshop personnel resources. This effectively solves the problems of relying on experience-based judgment for personnel scheduling in traditional production processes, which can easily lead to raw material backlog or equipment waiting for operation. It achieves a dynamic balance between personnel allocation and equipment load. Simultaneously, based on the updated personnel distribution and equipment load data after personnel scheduling, a dynamic visualization correlation diagram between personnel distribution and equipment load is generated. This visually presents the dynamic correspondence between changes in personnel distribution and equipment load in each process area on the monitoring interface of the supervision and control and data acquisition system, allowing managers to grasp the entire production process status in real time and clearly. This solves the information gap problem in traditional monitoring models, achieving clear visibility, controllability, and adjustability of the entire peanut processing process. It effectively reduces batch-to-batch product quality fluctuations, ensures production stability, and significantly improves the overall production efficiency and market competitiveness of the enterprise.

[0025] The following is a detailed description of each step in the visualized full-process monitoring method for peanut processing provided in the embodiments of this application: In one embodiment of this application, step S110 involves acquiring multi-angle image data of personnel in various process areas of the peanut processing workshop. Specifically, fixed high-definition industrial cameras can be selected. Depending on the size of each process area and the obstruction caused by equipment layout, 3-5 cameras are arranged for each core process area to form a multi-angle overlapping field of view. The cameras are all installed on pillars around the process area, and the complementary viewing angles are achieved by adjusting the installation angles. For example, in the baking process area, due to obstructions such as ovens and material racks, and the heat radiation from high-temperature baking, four cameras of this model are arranged, installed at the four corners of the area, with viewing angles set to 30° downward, 45° downward, 20° side, and 35° side, respectively, to achieve no blind spots in the coverage of the baking area's operating table, material loading and unloading points, and the work area around the oven. In the cleaning process area, the space is relatively open with only a small amount of moisture obstruction, so three cameras of this model are arranged in a triangular layout to cover the entire work area, including the cleaning pool, material conveying platform, and water treatment area. In the screening process area, due to the large amount of dust, additional anti-scratch and dust-proof optical protective covers can be installed on the camera lenses.

[0026] In one embodiment of this application, step S120, which determines the real-time location coordinates and dwell time of personnel in each process area based on multi-angle image data, further includes the following steps: Figure 2 As shown, the specific content is as follows: Step S210: Based on multi-angle image data, a target detection algorithm is used to determine the human body bounding boxes of personnel in each process area, and based on the human body bounding boxes, a pose estimation network is used to extract the coordinates of key points of each human body. Step S220: Determine the real-time position coordinates of personnel in each process area based on the key point coordinates of each human body, and determine the process area where each personnel is located based on the real-time position coordinates and the coordinates of each preset process area; Step S230: Obtain a continuous image frame sequence within the work area where each person is located, and determine the dwell time of each person in each work area based on the continuous image frame sequence.

[0027] Specifically, the YOLOv8m object detection model can be used to determine the bounding boxes of personnel in each work area. For pose estimation, the HRNet-W48 pose estimation network can be used, which can accurately extract the coordinates of key human points, including the head, neck, left and right shoulders, left and right elbows, left and right wrists, torso, left and right hips, left and right knees, and left and right ankles, to meet the feature requirements for personnel position calculation. Next, through camera calibration and multi-sensor fusion, the pixel coordinates of the key human points are converted into real-time position coordinates in the world coordinate system. Then, through spatial coordinate comparison, the accurate determination of the work area to which the personnel belong is achieved, solving the mapping problem between visual pixel coordinates and actual physical spatial coordinates, while ensuring the accuracy of personnel work area assignment. Once a person is determined to belong to a preset work area, image frames of that person within that work area are continuously acquired, forming a continuous image frame sequence within the same work area. The time interval for a single frame image is determined based on the camera's frame rate. Simultaneously, a threshold for accumulating dwell time is set for each process area (e.g., ≥90 seconds for inspection areas, ≥60 seconds for baking / cooling / packaging processes). Dwell time accumulation is triggered only when the actual duration of a continuous image frame sequence within a certain process area exceeds the corresponding threshold. Short dwell times below the threshold are not counted as valid dwell time. Dwell time is updated every 5 seconds. The actual dwell time = (end frame number of the continuous image frame sequence - start frame number) / camera frame rate.

[0028] For example, taking the actual monitoring process of the roasting process area in a peanut processing workshop as an example, multi-angle image data of the process area (resolution 1920×1080, frame rate 25fps) was acquired. A frame of the image was input into the YOLOv8m object detection model. The object detection model completed the detection and selected the human body bounding boxes of 4 workers within 16ms. After cropping the bounding boxes, they were input into the HRNet-W48 pose estimation network, which successfully extracted the coordinates of each person's key points. Subsequently, the pixel coordinates of the key points at the midpoint of the torso of the four individuals were converted using camera intrinsic and extrinsic parameters. Then, through Kalman filtering combined with a multi-camera spatiotemporal fusion algorithm, the real-time position coordinates of the four individuals in the world coordinate system were obtained as follows: (12.5, 8.3, 1.7), (13.2, 9.1, 1.7), (14.0, 8.5, 1.7), and (14.8, 9.3, 1.7) (unit: meters). Simultaneously, these real-time position coordinates were compared with the preset polygonal world coordinates (11.0, 7.0, 0), (16.0, 7.0, 0), (16.0, 10.0, 0), and (11.0, 10.0, 0) of the baking process area, confirming that all four individuals belonged to the baking process area. Next, four personnel were assigned unique IDs and continuously tracked. Personnel with ID 003 was determined to be in the baking process area. The continuous image frame sequence started at frame 1200. As of frame 5450, they had not left the area. Based on a frame rate of 25fps, their continuous dwell time was (5450-1200) / 25=170 seconds, which exceeded the 60-second cumulative trigger threshold of the baking process. The dwell time accumulation statistics were then triggered and updated every 5 seconds. At the 5th second, the continuous image frame sequence of this personnel reached frame 5575, and the dwell time was updated to 175 seconds.

[0029] In the above method, the precise and rapid extraction of human body bounding boxes and key point coordinates of personnel in each process area of ​​the peanut processing workshop is achieved through the collaborative application of target detection algorithm and pose estimation network. This solves the problem that traditional manual inspection cannot perform visual feature quantification extraction of personnel, and provides highly reliable visual basic data for personnel position coordinate calculation.

[0030] In one embodiment of this application, step S130 involves determining the personnel distribution in each process area based on real-time location coordinates and dwell time. The personnel distribution includes personnel gathering frequency and inter-personal collaboration distance. The steps also include the following: Figure 3 As shown, the specific content is as follows: Step S310: Based on the real-time location coordinates, use density clustering algorithm to determine the personnel cluster set for each process area, and based on the dwell time, count the total dwell time and the number of occurrences of each person in the personnel cluster set for each process area; Step S320: Determine the personnel gathering frequency in each process area based on the total dwell time and the number of occurrences, and determine the collaborative distance between personnel in each process area based on the personnel cluster set using the Euclidean distance method.

[0031] Specifically, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) density-based spatial clustering algorithm can be used for cluster analysis. This algorithm does not require pre-setting the number of clusters and can accurately identify high-density areas of people gathering in space. The engineering parameters are calibrated as a neighborhood radius ε = 8 meters and a minimum number of points MinPts = 5. That is, if there are at least 5 people within an 8-meter radius of a location, the area is considered a valid gathering area and included in the personnel cluster set; scattered individuals with fewer than this number are considered noise points. The total dwell time is the cumulative value of the effective dwell time of all individuals within a given personnel cluster set within a specified statistical period; the frequency of occurrence is the number of different individuals entering the personnel cluster set within the specified statistical period, i.e., the total number of people gathering in the cluster area per unit time. Personnel gathering frequency refers to the number of personnel entering each process area cluster within a unit of time. It is a core quantitative indicator characterizing the degree of personnel gathering in a process area. If there are multiple personnel clusters in a certain process area, the sum of the number of times each cluster appears within a unit of time is used to obtain the personnel gathering frequency of that process area.

[0032] For example, taking the roasting and packaging areas of a peanut processing workshop as examples, a statistical period of 1 hour is used. The density clustering algorithm parameters are set to a neighborhood radius ε = 8 meters, and the preset threshold for the collaborative distance between personnel is 6 meters. Real-time location coordinates of 12 personnel were collected in the roasting area. After inputting the data into the density clustering algorithm, cluster analysis was performed, resulting in one personnel cluster (containing 10 personnel) and 2 scattered noise points. Real-time location coordinates of 10 personnel were collected in the packaging area, and cluster analysis yielded one personnel cluster (containing 4 personnel) and 6 scattered noise points. Subsequently, it was found that the total dwell time of the 10 personnel in the roasting area cluster was 820 minutes, and the number of occurrences was 10. The total dwell time of the 4 personnel in the packaging area cluster was 150 minutes, and the number of occurrences was 4. Based on the above statistical data, the personnel gathering frequency was calculated. The baking process area had only one cluster, with a gathering frequency of 10 people / hour and an average dwell time of 820 / 10 = 82 minutes / person. The packaging process area had a gathering frequency of 4 people / hour and an average dwell time of 150 / 4 = 37.5 minutes / person. The Euclidean distance method was then used to calculate the collaborative distance between personnel. The real-time location coordinates of 10 personnel within the baking process cluster were extracted and paired, resulting in 45 coordinate pairs. The calculated collaborative distance set consisted of 45 distance values, with an average collaborative distance of 4.5 meters, and 40 distance values ​​less than 6 meters. The real-time location coordinates of 4 personnel within the packaging process cluster were extracted and paired, resulting in 6 coordinate pairs. The calculated collaborative distance set had an average collaborative distance of 10.8 meters, with only one distance value less than 6 meters. The final determined personnel gathering frequency in the baking process area was 10 people / hour and the working distance between personnel was 4.5 meters, while the personnel gathering frequency in the packaging process area was 4 people / hour and the working distance between personnel was 10.8 meters.

[0033] In the above methods, spatial clustering algorithms are used to accurately group personnel in each process area according to their spatial distribution characteristics, effectively distinguishing between personnel clusters and scattered noise points, and eliminating the interference of scattered personnel on the analysis of cluster status. By quantifying the frequency of personnel gathering, the industry pain point of traditional peanut processing workshops being unable to accurately quantify the personnel gathering status and relying solely on subjective human judgment is solved. By using the Euclidean distance method, the collaborative distance between personnel is accurately quantified, truly reflecting the spatial collaboration tightness of personnel gathered in each process area, and solving the problem of lack of quantitative data support for the collaborative status.

[0034] In one embodiment of this application, step S130, which determines the comprehensive density value of personnel in each process area based on the frequency of personnel gathering and the collaborative distance between personnel, further includes the following steps: Figure 4 As shown, the specific content is as follows: Step S410: Obtain the weights corresponding to the frequency of personnel gathering and the weights corresponding to the collaborative distance between personnel, respectively; Step S420: Based on the weight corresponding to the personnel gathering frequency and the weight corresponding to the personnel collaboration distance, perform a weighted calculation on the personnel gathering frequency and the personnel collaboration distance, and use the result of the weighted calculation as the comprehensive density value of personnel in each process area.

[0035] Specifically, taking the roasting and packaging areas of the peanut processing workshop as examples, the personnel gathering frequency in the roasting area is 10 people / hour, and the inter-personal collaboration distance is 4.5 meters. In the packaging area, the personnel gathering frequency is 4 people / hour, and the inter-personal collaboration distance is 10.8 meters. The weights corresponding to both personnel gathering frequency and inter-personal collaboration distance are 0.5. Simultaneously, the preset maximum personnel gathering frequency for all process areas in the entire workshop is 12 people / hour, and the maximum and minimum inter-personal collaboration distances are 12 meters and 3 meters, respectively. After standardization, the standardized personnel gathering frequency in the roasting area is 10 / 12 ≈ 0.833, and the standardized inter-personal collaboration distance is (12-4.5) / (12-3) = 0.833. The standardized personnel gathering frequency in the packaging area is 4 / 12 ≈ 0.333, and the standardized inter-personal collaboration distance is (12-10.8) / (12-3) ≈ 0.133. Therefore, the overall personnel density value in the baking process area is 0.5×0.833+0.5×0.833=0.833, and the overall personnel density value in the packaging process area is 0.5×0.333+0.5×0.133=0.233. The above two values ​​directly reflect that the personnel concentration in the baking process area is higher than that in the packaging process area.

[0036] In one embodiment of this application, step S140 involves obtaining the operating parameters of the equipment in each process area and determining the load index of the equipment in each process area based on the operating parameters. The load index includes the real-time deviation value of the machining cycle time, the fluctuation amplitude of the spindle load rate, and the peak intensity of the vibration spectrum characteristics. The operating parameters include the spindle current sequence and the vibration acceleration sequence. The process also includes the following steps: Figure 5 As shown, the specific content is as follows: Step S510: Within a preset time window, calculate the average current value in the spindle current sequence, and determine the real-time deviation value of the machining cycle based on the average current value and the preset reference current value. Step S520: Determine the spindle load rate sequence based on the spindle current sequence, and determine the spindle load rate fluctuation range based on the spindle load rate sequence; Step S530: Determine the characteristic peak intensity of the vibration spectrum based on the vibration acceleration sequence.

[0037] Specifically, the real-time deviation of the processing cycle time directly reflects the degree of deviation between the actual processing rhythm and the standard rhythm. The spindle current of peanut processing equipment is strongly positively correlated with the processing cycle time (the higher the current, the faster the processing cycle time; the more stable the current, the more standard the processing rhythm). Therefore, this deviation value can be calculated using current characteristics. For example, the spindle current sequence obtained by a current sensor at a sampling frequency of 100Hz can be slide-trimmed within a preset time window of 10 seconds. Each window corresponds to 1000 current data points. First, the current data within the window is denoised to filter out current interference caused by non-processing factors such as power grid fluctuations and equipment mechanical vibrations. Then, the arithmetic mean of all current data points within the denoised window is calculated to obtain the average spindle current value within the preset time window. The formula for calculating the real-time deviation of the processing cycle time is:

[0038] in, To determine the real-time deviation value of the processing cycle time, The average current value within a preset time window. This is the preset reference current value. The sign of this deviation value indicates the direction of the cycle time deviation: a positive result indicates that the current processing cycle time of the equipment is faster than the standard cycle time; a negative result indicates that the current processing cycle time of the equipment is slower than the standard cycle time; a result close to 0 indicates that the processing cycle time of the equipment is consistent with the standard cycle time.

[0039] Specifically, the spindle load rate fluctuation amplitude characterizes the dispersion of the spindle load over time, reflecting the stability of equipment operation. The greater the load rate fluctuation, the more unstable the equipment operation and the more prone the load is to abnormal conditions. This indicator is also based on the strong positive correlation between spindle current and load rate. The spindle current sequence can be denoised, and then segmented into 1-second intervals, with each segment corresponding to 100 current data points. The effective value of each segment is calculated and directly used as the spindle load rate for the current second. The spindle load rates for each second are then arranged chronologically to form a continuous spindle load rate sequence. For example, the spindle load rate sequence can be truncated in 30-second windows, with each window corresponding to 30 spindle load rate data points. The standard deviation of all data points within the window is calculated; this standard deviation represents the spindle load rate fluctuation amplitude at the current moment. The larger the standard deviation, the more drastic the change in spindle load rate within the window, and the worse the equipment's operational stability.

[0040] Specifically, the peak intensity of the vibration spectrum characteristic can characterize the intensity of vibration in the mechanical structure of equipment. Wear and failure of core components such as the main shaft and bearings of peanut processing equipment will be reflected in the peak changes of the vibration spectrum characteristic. Therefore, this indicator can reflect the load status from the perspective of equipment health, avoiding excessive load or equipment damage caused by abnormal vibration. The vibration acceleration sequence obtained by the accelerometer at a sampling frequency of 100Hz can be truncated into a fixed window of 1024 data points. To reduce spectral leakage, the vibration acceleration data within each window is windowed using a Hanning window. The windowed vibration acceleration data is then subjected to a Fast Fourier Transform to convert the time-domain vibration acceleration sequence into frequency-domain vibration spectrum data, obtaining the vibration spectrum characteristic of the equipment. This characteristic, with frequency as the abscissa and vibration acceleration amplitude as the ordinate, can clearly reflect the vibration intensity of the equipment at different frequencies. For example, vibration faults in peanut processing equipment are mostly concentrated at the spindle frequency and its 2nd and 3rd harmonic frequencies. Therefore, the system extracts the peak amplitude of vibration acceleration at the three characteristic frequency points of the fundamental frequency, 2nd harmonic frequency, and 3rd harmonic frequency for the preset spindle frequency of each process equipment. Then, according to the degree of influence of each frequency point on the health of the equipment, weighting coefficients are assigned, and the characteristic peak intensity of the vibration spectrum is obtained by weighted summation.

[0041] For example, taking the roasting machine in the peanut processing workshop as a monitoring example, the equipment has a preset reference current value of 10A and a preset spindle frequency of 50Hz. The accelerometer collects the spindle current sequence and vibration acceleration sequence at a sampling frequency of 100Hz. After extracting the spindle current sequence within a preset time window of 10 seconds and denoising it, the average current value is calculated to be 10.78A. Substituting this into the above formula, the real-time deviation of the processing cycle is calculated to be (10.78-10) / 10×100%=+7.8%, indicating that the current processing cycle of the roasting machine is 7.8% faster than the standard cycle. The effective current value is calculated for the denoised spindle current sequence in 1-second segments, generating a spindle load rate sequence such as [78%, 79%, 81%, 77%, 82%, ...]. The spindle load rate fluctuation is calculated to be 12.4% with a 30-second moving window. The vibration acceleration sequence was truncated into a window of 1024 data points and windowed using a Hanning window. After Fast Fourier Transform (FFT), the vibration spectrum characteristics were obtained. The peak amplitudes at 50Hz (fundamental frequency), 100Hz (second harmonic), and 150Hz (third harmonic) were extracted and found to be 0.035g, 0.048g, and 0.021g, respectively. Weighted by 0.4 for the fundamental frequency, 0.4 for the second harmonic, and 0.2 for the third harmonic, the peak intensity of the vibration spectrum characteristic was calculated to be 0.4×0.035 + 0.4×0.048 + 0.2×0.021 = 0.0374g. Finally, through the above calculations, the load indicators of the baking machine were obtained as follows: real-time deviation of processing cycle time +7.8%, spindle load rate fluctuation amplitude 12.4%, and peak intensity of vibration spectrum characteristic 0.0374g.

[0042] In one embodiment of this application, step S150, which determines the comprehensive load value of the equipment in each process area based on the load index, further includes the following steps: Figure 6 As shown, the specific content is as follows: Step S610: Obtain the weights corresponding to the real-time deviation value of the machining cycle time, the weights corresponding to the spindle load rate fluctuation amplitude, and the weights corresponding to the peak intensity of the vibration spectrum characteristics, respectively. Step S620: Based on the weights corresponding to the real-time deviation value of the machining cycle time, the weights corresponding to the fluctuation amplitude of the spindle load rate, and the weights corresponding to the peak intensity of the vibration spectrum characteristics, perform a weighted calculation on the real-time deviation value of the machining cycle time, the fluctuation amplitude of the spindle load rate, and the peak intensity of the vibration spectrum characteristics, and use the result of the weighted calculation as the comprehensive load value of the equipment in each process area.

[0043] Specifically, for example, the weight corresponding to the real-time deviation of the processing cycle time is 0.4, the weight corresponding to the spindle load rate fluctuation is 0.3, and the weight corresponding to the peak intensity of the vibration spectrum is 0.3. The real-time deviation of the processing cycle time is +7.8%, the spindle load rate fluctuation is 12.4%, and the peak intensity of the vibration spectrum is 0.0374g. The extreme values ​​of the corresponding indicators for similar processing equipment in the entire workshop are: maximum real-time deviation of processing cycle time (20% maximum, -15% minimum), maximum spindle load rate fluctuation (20% maximum, 2% minimum), and maximum peak intensity of vibration spectrum (0.08g maximum, 0.01g minimum). Subsequently, the three indicators are linearly normalized to 0-1, resulting in a standardized real-time deviation of processing cycle time of 0.651, a spindle load rate fluctuation of 0.578, and a peak intensity of vibration spectrum of 0.391. Finally, a weighted calculation is performed to obtain the comprehensive load value of the baking machine: 0.4 × 0.651 + 0.3 × 0.578 + 0.3 × 0.391 = 0.5511.

[0044] In one embodiment of this application, in step S160, if the deviation between the comprehensive density value and the comprehensive load value exceeds a preset deviation threshold, the current process area is marked as an overloaded area; if the deviation between the comprehensive density value and the comprehensive load value does not exceed the preset deviation threshold, the current process area is marked as an underloaded area, and a list of personnel to be transferred from the overloaded area to the underloaded area is determined. The process also includes the following steps: Figure 7 As shown, the specific content is as follows: Step S710: Obtain the movement trajectory vectors of personnel within the overloaded area; Step S720: Based on the movement trajectory vector, use the trajectory similarity clustering method to determine the trajectory clusters of each person, and count the task switching frequency of each person's trajectory cluster within a preset time period; Step S730: If the task switching frequency does not exceed the preset task switching frequency, the personnel corresponding to the current personnel trajectory cluster are taken as candidates, and the continuous working time of each candidate is obtained. Step S740: If the continuous working time does not exceed the preset continuous working time threshold, the current candidate personnel will be used as the transfer personnel, and the list of personnel to be transferred from the overloaded area to the underloaded area will be determined based on the transfer personnel.

[0045] Specifically, if the deviation between the overall density value and the overall load value exceeds the preset deviation threshold, it indicates that the personnel density in the process area is too high, exceeding the actual load-bearing requirements of the equipment, and the current process area is marked as an overloaded area; if the deviation between the overall density value and the overall load value does not exceed the preset deviation threshold, it indicates that the personnel density in the process area is insufficient and cannot match the personnel requirements for high-load operation of the equipment, and the current process area is marked as an underloaded area.

[0046] Specifically, personnel movement trajectory vectors can accurately represent standardized data of personnel movement characteristics in both spatial and temporal dimensions within an overloaded area. The data source for this vector is the fusion of visual tracking data from the workshop camera network and data collected by positioning devices worn by personnel. For example, a unique ID is assigned to all on-duty personnel within the overloaded area, and then the complete original movement data of each person within a preset time period is extracted and converted into a standardized personnel movement trajectory vector. This vector, with time as the horizontal axis and spatial location as the vertical axis, is composed of multiple ordered spatiotemporal nodes, clearly reflecting the personnel's movement range, work area preferences, and movement patterns between workstations.

[0047] Specifically, Dynamic Time Warping (DTW) can be used for trajectory similarity clustering analysis. This algorithm can effectively handle time-series trajectory data in industrial scenarios, ignoring minor differences in trajectory over time and accurately identifying personnel with similar spatial movement characteristics. A trajectory similarity threshold can be preset, grouping personnel whose trajectory similarity exceeds the preset threshold into the same personnel trajectory cluster. Trajectory similarity implies a high degree of overlap in personnel movement range and work areas, resulting in lower work adaptation costs when transferring to adjacent underloaded areas. If the personnel trajectory similarity does not exceed the preset threshold, it is marked as a scattered trajectory personnel and not included in subsequent screening. Secondly, for each personnel trajectory cluster obtained from clustering, its task switching frequency within a preset time period is calculated. The task switching frequency refers to the arithmetic mean of the number of task switching times between different workstations / sub-processes within the overloaded area for personnel in each personnel trajectory cluster, expressed as times per hour. Task switching behavior can be determined by the positional changes of personnel trajectory vectors and the results of work action recognition (e.g., moving from the loading station to the unloading station in the baking process and completing the work action change is counted as one task switch). The average number of switches for each cluster is calculated on an hourly basis to obtain the task switching frequency for each personnel trajectory cluster. The preset time period can be based on the regular production shift system of the peanut processing workshop (8-hour single shift / 12-hour double shift) and the complete work cycle of each process (including raw material arrival, processing, semi-finished product flow, and finished product temporary storage), ensuring that the extracted personnel movement trajectory vectors can cover the complete work patterns of personnel and avoid deviations in trajectory similarity clustering due to insufficient data. For example, during peak production periods or when the processing pace of processes is accelerated, the preset time period for task switching frequency statistics can be shortened to 30 minutes to improve the real-time performance of personnel status statistics and adapt to rapid scheduling needs; during off-peak production periods (when the overall equipment load value is below 0.3) or when the processing pace of processes is slowed down, the preset time period for trajectory extraction can be extended to 48 hours to further improve the accuracy of trajectory clustering.

[0048] Specifically, if the task switching frequency does not exceed the preset task switching frequency, it indicates that the work status of personnel within the current personnel trajectory cluster is stable and will not cause work process interruption due to frequent workstation / sub-process switching. Reassigning such personnel will not significantly affect the basic production in the overloaded area. If the task switching frequency of a personnel trajectory cluster exceeds the preset task switching frequency, it indicates that the work status of personnel within the trajectory cluster fluctuates greatly. Direct reassignment would disrupt the original work rhythm in the overloaded area. Therefore, such trajectory clusters are excluded, and only all personnel corresponding to trajectory clusters with task switching frequencies meeting the requirements are marked as candidate personnel. Subsequently, through the workshop's personnel attendance management system and continuous recording of work status data via camera visual recognition, the real-time continuous working time of each candidate personnel is accurately obtained. This time refers to the uninterrupted working time of the candidate personnel from the start of the current round of continuous work to the statistical time, in hours. Specifically, historical work data from each process in the peanut processing workshop over the past 3-6 months can be collected to calculate the average task switching frequency under normal working conditions. 1.2 times the average is taken as the preset task switching frequency for that process. For example, core processing steps such as baking and cooling involve fixed sub-stations such as loading, processing, and unloading, and the work process is relatively stable. The historical average frequency of preset task switching is about 3.3 times / hour, and the preset task switching frequency is 4 times / hour. Cleaning and grading processes have simple work processes and few sub-stations. The historical average frequency of preset task switching is about 1.7 times / hour, and the preset task switching frequency is 2 times / hour.

[0049] Specifically, if a candidate's continuous working time does not exceed the preset continuous working time threshold, it indicates that the candidate is in good physical condition and can handle the workload of cross-regional operations, and this candidate is officially marked as a seconded worker. If a candidate's continuous working time exceeds the preset continuous working time threshold, it indicates that the candidate is in a state of fatigue, and they are excluded from the secondment pool to ensure the safety and efficiency of production operations. Secondly, based on the actual number of personnel needed in the underloaded area and the skill tag matching degree between the seconded personnel and the underloaded area's processes, all seconded personnel can be prioritized, with priority given to personnel whose skills have a high matching degree with the underloaded area's process requirements. Then, the final number of seconded personnel is determined based on the number of personnel needed in the underloaded area, ultimately generating a standardized seconded personnel list. The seconded personnel list can include core information such as the personnel's unique ID, name, current workstation in the overloaded area, and target underloaded area process / workstation to be transferred to. This list can also be simultaneously synchronized to the peanut processing workshop's monitoring system, dispatch terminal, and on-site operation dashboard to ensure the efficient transmission of dispatch instructions. Among them, by regulating the continuous working hours of manual laborers, and taking into account the industry characteristics of peanut processing workshops which are mainly engaged in light to moderate physical labor, the preset continuous working hours can be set at 4 hours, which is in line with the work rest system of industrial production and ensures the work efficiency and physical condition of personnel.

[0050] For example, consider the roasting area in a peanut processing workshop as an overloaded area and the cooling area as an adjacent underloaded area. The roasting area has 20 workers, while the cooling area requires 5. The preset timeframe for trajectory extraction is 24 hours, the trajectory similarity threshold for the dynamic time warping algorithm is ≥0.8, the preset task switching frequency is 4 times / hour, and the preset continuous working time threshold is 4 hours. Unique IDs are assigned to the 20 workers in the roasting area. Visual tracking and positioning data for each worker are extracted over 24 hours and converted into standardized personnel movement trajectory vectors. These vectors clearly reflect the movement patterns of workers at the three core workstations of the roasting process: loading, roasting, and unloading. Simultaneously, the dynamic time warping algorithm is used to cluster the movement trajectory vectors of the 20 workers, resulting in 3 personnel trajectory clusters and 0 individuals with scattered trajectories. Personnel trajectory cluster 1 includes 8 people, personnel trajectory cluster 2 includes 7 people, and personnel trajectory cluster 3 includes 5 people. Furthermore, the task switching frequencies for each personnel trajectory cluster were calculated as follows: 3 times / hour for trajectory cluster 1, 5 times / hour for trajectory cluster 2, and 2.5 times / hour for trajectory cluster 3. Trajectory cluster 2, with a task switching frequency of 5 times / hour, exceeded the preset threshold of 4 times / hour and was therefore excluded. The 13 personnel from trajectory clusters 1 and 3 were marked as candidates. Simultaneously, the real-time continuous working time of these 13 candidates was obtained. The working time for the 8 personnel in trajectory cluster 1 was 2.0-3.5 hours, and the working time for the 5 personnel in trajectory cluster 3 was 1.0-4.0 hours, both representing the real-time continuous working time for the current round. Since the continuous working time of all 13 candidates did not exceed the preset continuous working time threshold of 4 hours, they were all marked as seconded personnel. Based on the skill requirements of the cooling process area, the 13 transferred personnel were prioritized according to their skill matching degree. The top 5 personnel with the highest skill matching degree (ID001, ID003, ID006, ID012, ID018) were selected first. The final list of transferred personnel to be transferred from the overloaded baking area to the underloaded cooling area was determined, including the ID, name, original baking station and target cooling station of the 5 people, and was synchronized to the workshop monitoring system and on-site operation board.

[0051] In one embodiment of this application, step S170, which updates the personnel distribution and load index according to the list of transferred personnel, and determines a dynamic visual correlation diagram between personnel distribution and equipment load based on the updated personnel distribution and load index, further includes the following steps: Figure 8 As shown, the specific content is as follows: Step S810: Based on the updated personnel distribution, generate a personnel cluster heatmap using a heatmap rendering method; Step S820: Based on the updated load indicators, generate a sequence of device status icons using a visualization rendering method; Step S830: Determine the dynamic visualization correlation diagram between personnel distribution and equipment load based on the personnel gathering heat map and equipment status icon sequence.

[0052] Specifically, the updated personnel distribution refers to the complete personnel status data, including real-time personnel location coordinates, personnel gathering frequency, inter-personal collaboration distance, and overall personnel density, recalculated for each process area after personnel transfer based on the personnel reassignment list. The heatmap rendering method uses a pre-defined world coordinate system layout of the peanut processing workshop as the base map, and the overall personnel density value of each process area as the core rendering basis. It is a standardized visualization generation method that intuitively represents the density of personnel gathering and distribution through color gradients and brightness changes. Rendering rules can be pre-defined as follows: cool colors (blue, light blue) represent low overall personnel density and sparse personnel distribution; neutral colors (green, yellow) represent moderate overall personnel density and reasonable personnel distribution; and warm colors (orange, red) represent high overall personnel density and dense personnel gathering. The rendering range fully covers all process areas, including cleaning, grading, baking, cooling, screening, and packaging, as well as auxiliary operation areas. The rendering results are precisely bound to the real-time personnel location coordinates and can be dynamically refreshed in milliseconds as the personnel distribution data is updated, ultimately generating a personnel gathering heatmap that intuitively reflects the real-time gathering status of personnel in each area.

[0053] Specifically, the updated load indicators refer to the complete equipment operation data, including the real-time deviation of processing cycle time, spindle load rate fluctuation amplitude, peak intensity of vibration spectrum characteristics, and comprehensive equipment load value, recalculated after personnel reassignment and load rebalancing. Visualization rendering is an industrial visualization rendering method that matches standardized graphics, colors, and icons to different equipment operating states and load levels. For example, corresponding basic equipment icons can be designed based on the equipment type of each process (baking equipment, packaging equipment, cleaning equipment, etc.). Then, based on the comprehensive equipment load value and individual load indicators, status rendering rules are set as follows: green represents normal equipment load and stable operation; yellow represents slightly abnormal equipment load and is in a warning state; red represents severely abnormal equipment load and requires timely intervention. The equipment icons for each process area are arranged sequentially according to the actual equipment layout in the workshop. Each equipment icon is synchronously associated with the corresponding real-time load indicator value, forming a sequence of equipment status icons that is completely consistent with the workshop equipment layout and can be dynamically updated with the load indicators.

[0054] Specifically, using a unified global coordinate system layout of the peanut processing workshop as a benchmark, the generated personnel gathering heatmap and equipment status icon sequence can be spatially precisely aligned and layered, ensuring complete overlap between personnel gathering areas and corresponding process equipment locations, with real-time data linkage. This creates a linked and visualized screen integrating personnel distribution and equipment load status—a dynamic and visualized correlation diagram between personnel distribution and equipment load. Managers can simultaneously observe the personnel density and corresponding equipment load level of a specific process area through a single interface, intuitively grasping the matching relationship between the two. Furthermore, all screen elements are updated in real-time as personnel move and equipment operating status changes, achieving dynamic and visualized monitoring of the entire process. This approach achieves an integrated, visualized, and dynamic presentation of the personnel distribution and equipment load status in the peanut processing workshop. It solves the information gap problem in the traditional monitoring model, where personnel distribution and equipment load are disconnected and their matching relationship cannot be intuitively observed. The visualized screen closely matches the actual layout of the workshop, the data is accurately linked, and the rendering rules conform to the usage habits of industrial monitoring. This helps managers quickly and intuitively grasp the human-machine collaboration status and load balance effect of each process in the entire workshop, greatly improving the efficiency and accuracy of production monitoring and scheduling decisions. At the same time, the dynamic refresh mechanism ensures the real-time nature of the visualized information, providing an intuitive and reliable display carrier and decision support for the visualized, intelligent, and refined management and control of the entire peanut processing process.

[0055] This application also provides a visualized full-process monitoring system for peanut processing. This system may include a first acquisition module, a first determination module, a second determination module, a second acquisition module, a third determination module, a fourth determination module, and a fifth determination module. Specifically, the first acquisition module acquires multi-angle image data of personnel in each process area of ​​the peanut processing workshop; the first determination module determines the real-time location coordinates and dwell time of personnel in each process area based on the multi-angle image data; the second determination module determines the personnel distribution in each process area based on the real-time location coordinates and dwell time, including personnel gathering frequency and inter-personal cooperation distance, and determines the comprehensive personnel density value in each process area based on the personnel gathering frequency and inter-personal cooperation distance; the second acquisition module acquires the operating parameters of the equipment in each process area and determines the load index of the equipment in each process area based on the operating parameters, including the real-time deviation of the processing cycle time and the spindle load rate. The third determination module is used to determine the comprehensive load value of equipment in each process area based on the load index, including fluctuation amplitude and peak intensity of vibration spectrum characteristics. The fourth determination module is used to mark the current process area as an overload area if the deviation between the comprehensive density value and the comprehensive load value exceeds a preset deviation threshold, and to mark the current process area as an underload area if the deviation between the comprehensive density value and the comprehensive load value exceeds a preset deviation threshold, and to determine the list of personnel to be transferred from the overload area to the underload area. The fifth determination module is used to update the personnel distribution and load index based on the list of personnel to be transferred, and to determine the dynamic visualization correlation diagram between personnel distribution and equipment load based on the updated personnel distribution and load index.

[0056] It should be noted that the embodiments of the peanut processing visualization full-process monitoring system provided in this application can be used to execute the processing flow of the embodiments of the peanut processing visualization full-process monitoring method described above. Its functions will not be repeated here, but can be found in the detailed description of the above method embodiments. All modules of this system are integrated into the hardware architecture of the supervision and control and data acquisition system, achieving coordinated implementation of software and hardware.

[0057] This application also provides an electronic device, which includes one or more processors and memory resources represented by a memory for storing instructions executable by the processor, such as application programs. The application programs stored in the memory may include one or more modules, each corresponding to a set of instructions. Furthermore, the processor is configured to execute instructions to perform the aforementioned method for visually monitoring the entire peanut processing process.

[0058] The electronic device may also include a power supply component configured to perform power management of the electronic device, a wired or wireless network interface configured to connect the electronic device to a network, and an input / output (I / O) interface. The electronic device can be operated based on operating devices stored in memory, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or similar.

[0059] In one embodiment, a computer device, which may be a server, is also provided. The computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device stores data. The I / O interfaces of the computer device are used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a visualized, end-to-end monitoring method for peanut processing.

[0060] In one embodiment, a computer device is provided, which may be a terminal. The computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for visualized full-process monitoring of peanut processing. The display unit of the computer device is used to form a visually visible image and may be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0061] This application may take the form of a computer program product implemented on one or more storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing program code. Computer-readable storage media include permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information may be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to: phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0062] It should be noted that although the steps of the visualized full-process monitoring method for peanut processing in this application are described in a specific order in the accompanying drawings, this does not require or imply that these steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps, such as omitting certain steps, combining multiple steps into one step, and / or breaking down one step into multiple steps, should all be considered part of this application.

[0063] It should be understood that this application is not limited to the detailed structure and arrangement of the modules in the visualized full-process monitoring system for peanut processing proposed in this specification. This application can have other implementations and can be implemented and executed in various ways. The foregoing variations and modifications fall within the scope of this application. It should be understood that the application and its definition in this specification extend to all alternative combinations of two or more individual features mentioned or apparent in the text and / or drawings. All these different combinations constitute multiple alternative aspects of this application. The embodiments described in this specification illustrate the best known mode for implementing this application and will enable those skilled in the art to utilize this application.

Claims

1. A method for visualized full-process monitoring of peanut processing, characterized in that, include: Acquire multi-angle image data of personnel in various process areas of the peanut processing workshop; Based on the multi-angle image data, the real-time location coordinates and dwell time statistics of personnel in each process area are determined; The personnel distribution in each process area is determined based on the real-time location coordinates and the dwell time statistics. The personnel distribution includes the personnel gathering frequency and the cooperation distance between personnel. The comprehensive personnel density value in each process area is determined based on the personnel gathering frequency and the cooperation distance between personnel. The operating parameters of the equipment in each process area are obtained, and the load index of the equipment in each process area is determined based on the operating parameters. The load index includes the real-time deviation of the machining cycle time, the fluctuation amplitude of the spindle load rate, and the peak intensity of the vibration spectrum characteristics. The comprehensive load value of the equipment in each process area is determined based on the aforementioned load index. If the deviation between the comprehensive density value and the comprehensive load value exceeds a preset deviation threshold, the current process area is marked as an overloaded area; if the deviation between the comprehensive density value and the comprehensive load value does not exceed the preset deviation threshold, the current process area is marked as an underloaded area, and a list of personnel to be transferred from the overloaded area to the underloaded area is determined. The personnel distribution and load index are updated based on the list of personnel to be transferred, and a dynamic visualization correlation diagram between personnel distribution and equipment load is determined based on the updated personnel distribution and load index.

2. The method for visually monitoring the entire peanut processing process according to claim 1, characterized in that, The step of determining the real-time location coordinates and dwell time of personnel in each process area based on the multi-angle image data includes: Based on the multi-angle image data, a target detection algorithm is used to determine the human body bounding boxes of personnel in each process area, and based on the human body bounding boxes, a pose estimation network is used to extract the coordinates of key points of each human body. The real-time position coordinates of personnel in each process area are determined based on the coordinates of each key human body point, and the process area where each personnel is located is determined based on the real-time position coordinates and the coordinates of each preset process area. Obtain a continuous sequence of image frames within the work area where each person is located, and determine the dwell time of each person in each work area based on the continuous sequence of image frames.

3. The method for visually monitoring the entire peanut processing process according to claim 1, characterized in that, The step of determining the personnel distribution in each process area based on the real-time location coordinates and the dwell time includes: Based on the real-time location coordinates, a density clustering algorithm is used to determine the personnel cluster set for each process area, and based on the dwell time, the total dwell time and the number of occurrences of each person in the personnel cluster set for each process area are counted. The frequency of personnel gathering in each process area is determined based on the total dwell time and the number of occurrences. Based on the personnel cluster set, the Euclidean distance method is used to determine the collaborative distance between personnel in each process area.

4. The method for visually monitoring the entire peanut processing process according to claim 1, characterized in that, The determination of the overall personnel density value for each process area based on the personnel gathering frequency and the inter-personal cooperation distance includes: Obtain the weights corresponding to the frequency of personnel gathering and the weights corresponding to the interpersonal collaboration distance, respectively. Based on the weights corresponding to the personnel gathering frequency and the personnel collaboration distance, a weighted calculation is performed on the personnel gathering frequency and the personnel collaboration distance, and the result of the weighted calculation is used as the comprehensive density value of personnel in each process area.

5. The method for visually monitoring the entire peanut processing process according to claim 1, characterized in that, The step of determining the load index of equipment in each process area based on the operating parameters includes: The operating parameters include the spindle current sequence and the vibration acceleration sequence; Within a preset time window, the average current value in the spindle current sequence is calculated, and the real-time deviation value of the machining cycle time is determined based on the average current value and the preset reference current value. The spindle load rate sequence is determined based on the spindle current sequence, and the spindle load rate fluctuation range is determined based on the spindle load rate sequence. The peak intensity of the vibration spectrum is determined based on the vibration acceleration sequence.

6. The method for visualized full-process monitoring of peanut processing according to claim 1, characterized in that, The determination of the comprehensive load value of equipment in each process area based on the load index includes: The weights corresponding to the real-time deviation value of the machining cycle time, the weights corresponding to the fluctuation amplitude of the spindle load rate, and the weights corresponding to the peak intensity of the vibration spectrum characteristics are obtained respectively. Based on the weights corresponding to the real-time deviation value of the machining cycle time, the weights corresponding to the fluctuation amplitude of the spindle load rate, and the weights corresponding to the peak intensity of the vibration spectrum characteristics, a weighted calculation is performed on the real-time deviation value of the machining cycle time, the fluctuation amplitude of the spindle load rate, and the peak intensity of the vibration spectrum characteristics, and the result of the weighted calculation is used as the comprehensive load value of the equipment in each process area.

7. The method for visualized full-process monitoring of peanut processing according to claim 1, characterized in that, The list of personnel to be transferred from the overloaded area to the underloaded area includes: Obtain the movement trajectory vectors of personnel within the overloaded area; Based on the movement trajectory vector, the trajectory similarity clustering method is used to determine the trajectory clusters of each person, and the task switching frequency of each trajectory cluster within a preset time period is statistically analyzed. If the task switching frequency does not exceed the preset task switching frequency, then the personnel corresponding to the current personnel trajectory cluster are taken as candidates, and the continuous working time of each candidate is obtained. If the continuous working time does not exceed the preset continuous working time threshold, the current candidate personnel will be used as transfer personnel, and a list of transfer personnel to be transferred from the overloaded area to the underloaded area will be determined based on the transfer personnel.

8. The method for visually monitoring the entire peanut processing process according to claim 1, characterized in that, The step of determining a dynamic visual correlation diagram between personnel distribution and equipment load based on the updated personnel distribution and load index includes: Based on the updated personnel distribution, a heatmap rendering method is used to generate a personnel cluster heatmap. Based on the updated load index, a sequence of device status icons is generated using a visualization rendering method. A dynamic visual correlation diagram between personnel distribution and equipment load is determined based on the personnel gathering heat map and the equipment status icon sequence.

9. A visualized end-to-end monitoring system for peanut processing, characterized in that, include: The first acquisition module is used to acquire multi-angle image data of personnel in various process areas of the peanut processing workshop; The first determining module is used to determine the real-time location coordinates and dwell time of personnel in each process area based on the multi-angle image data; The second determining module is used to determine the personnel distribution in each process area based on the real-time location coordinates and the dwell time. The personnel distribution includes personnel gathering frequency and personnel cooperation distance, and determines the comprehensive density value of personnel in each process area based on the personnel gathering frequency and personnel cooperation distance. The second acquisition module is used to acquire the operating parameters of the equipment in each process area, and determine the load index of the equipment in each process area based on the operating parameters. The load index includes the real-time deviation value of the processing cycle, the fluctuation amplitude of the spindle load rate, and the peak intensity of the vibration spectrum characteristics. The third determining module is used to determine the comprehensive load value of the equipment in each process area based on the load index; The fourth determining module is used to mark the current process area as an overloaded area if the deviation between the comprehensive density value and the comprehensive load value exceeds a preset deviation threshold, and to mark the current process area as an underloaded area if the deviation between the comprehensive density value and the comprehensive load value exceeds the preset deviation threshold, and to determine the list of personnel to be transferred from the overloaded area to the underloaded area. The fifth determining module is used to update the personnel distribution and the load index according to the list of personnel to be transferred, and to determine a dynamic visual correlation diagram between personnel distribution and equipment load according to the updated personnel distribution and the load index.

Images