Terminal self-unloading vehicle operation mode optimization method and system based on differential box algorithm

Abstract

The present application relates to the technical field of port intelligence, and discloses a wharf self-unloading vehicle operation mode optimization method and system based on a differential box algorithm, which collects self-unloading vehicle operation process images, calculates fractal dimension features by using the differential box algorithm, and establishes a correlation model of fractal dimension and energy consumption data. Firstly, a plurality of groups of operation images are preprocessed, fractal dimension features are extracted, and a feature database is constructed; the fractal dimension threshold interval of a high energy consumption cluster is determined through cluster analysis; after the fractal dimension of the operation image is calculated in real time, it is compared with the threshold interval of the high energy consumption cluster, and the operation mode is dynamically adjusted according to the comparison result, so that the real-time optimization of the operation mode is realized, and the energy consumption efficiency is significantly improved and the operation cost is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of port intelligence, and particularly relates to a wharf self-unloading truck operation mode optimization method and system based on a differential box algorithm. BACKGROUND

[0002] With the rapid development of global trade, port logistics operation efficiency has become a key factor affecting the effectiveness of the supply chain. Self-unloading trucks, especially wharf self-unloading trucks used for transporting bulk goods such as coal and ore, as the main horizontal transportation tool of bulk cargo wharves, the energy consumption management in the operation process is directly related to the operation cost and environmental performance of the port. At present, most ports mainly aim to improve the operation efficiency in the production process, and lack of fine energy consumption analysis means for the operation process of self-unloading trucks, resulting in low energy utilization efficiency and high fossil fuel consumption.

[0003] The traditional energy consumption monitoring method mainly relies on sensor data and artificial experience judgment, which has obvious limitations. On the one hand, the monitoring scheme based on physical sensors has high implementation cost and complex maintenance, and is difficult to be deployed on a large scale; on the other hand, the existing image processing technology is mainly concentrated in basic applications such as vehicle identification and track tracking, and cannot deeply mine the internal relationship between visual features and energy consumption indicators. These technical defects make it difficult for port managers to accurately assess the influence of different operation modes on energy consumption, and it is even more difficult to develop scientific and effective optimization strategies.

[0004] Fractal geometry theory, as an important tool for describing complex systems, has shown potential in engineering applications, but there are still obvious deficiencies in its application to equipment energy consumption analysis. In the calculation of fractal dimension by differential box algorithm, the parameter setting of differential box algorithm, as a mature fractal dimension calculation method, is crucial to the accuracy of the result. However, the current parameter selection mainly depends on artificial experience or simple heuristic rules, and lacks a systematic optimization method, which is difficult to adapt to the feature extraction requirements in different scenarios.

[0005] Therefore, there is an urgent need for a wharf self-unloading truck operation mode optimization method and system based on a differential box algorithm, which obtains the fractal dimension of the self-unloading truck operation mode picture by the differential box algorithm, and then determines the energy consumption of different vehicle operating states to optimize the operation mode. SUMMARY

[0006] In order to solve the above technical problems, the present application provides a wharf self-unloading truck operation mode optimization method and system based on a differential box algorithm, which obtains the fractal dimension of the self-unloading truck operation mode picture by the differential box algorithm, and determines the energy consumption of different vehicle operating states to optimize the operation mode.

[0007] The present application provides a wharf self-unloading truck operation mode optimization method based on a differential box algorithm, which comprises the following steps:

[0008] Step 1, collect multiple groups of dump truck running process pictures and pre-process; the dump truck running process picture is an MxM pixel picture;

[0009] Step 2, based on the multiple groups of pre-processed dump truck running process pictures, calculate the fractal dimension using an improved differential box algorithm;

[0010] Step 3, based on the multiple groups of pre-processed dump truck running process pictures, construct a feature database, the feature database includes: the fractal dimension corresponding to each dump truck running process picture and the corresponding energy consumption data;

[0011] Step 4, use K-means clustering analysis to identify the relationship between fractal dimension and energy consumption data, and set the fractal dimension threshold interval of the high energy consumption cluster according to the relationship;

[0012] Step 5, calculate the real-time fractal dimension according to the real-time collected dump truck running process picture;

[0013] Step 6, compare the real-time fractal dimension with the fractal dimension threshold interval of the high energy consumption cluster, and adjust the dump truck running mode according to the comparison result.

[0014] Further, in step 1, the pre-processing includes: gray processing, image size standardization and image enhancement processing on the multiple groups of dump truck running process pictures.

[0015] Further, in step 2, the improved differential box algorithm is: using a genetic algorithm to obtain the optimal scale parameter combination of the differential box algorithm.

[0016] Further, the use of a genetic algorithm to obtain the optimal scale parameter combination of the differential box algorithm specifically includes:

[0017] Step 2-1, randomly generate multiple scale parameter combinations as initial population, each scale parameter combination contains several different scale parameters;

[0018] Step 2-2, for each scale parameter combination in the population, calculate the fractal dimension sequence of the multiple sample images using the differential box algorithm, and calculate the fitness value based on the standard deviation of the fractal dimension and the contour coefficient in different energy consumption data intervals;

[0019] Step 2-3, select the next generation population according to the fitness value and the preset proportion;

[0020] Step 2-4, perform crossover and mutation operations on the selected next generation population to generate new scale parameter combinations;

[0021] Step 2-5, repeat steps 2-2 to 2-4 until the maximum number of iterations is reached, and output the optimal scale parameter combination.

[0022] Further, in step 2, the improved differential box algorithm is used to calculate the fractal dimension, specifically comprising:

[0023] Step 2-1-1, divide the pre-processed dump truck running process picture into a plurality of grids with size s i ×s i pixels, where s i is a scale parameter, and the value range of s i satisfies 2≤s i ≤M / 2, M is the width of the image;

[0024] Step 2-1-2, for each grid, calculate the maximum and minimum gray values of the pixels within it;

[0025] Step 2-1-3, determine the box height k based on the total gray level G of the image;

[0026] Step 2-1-4, calculate the number of boxes required to cover each grid according to the maximum and minimum gray values;

[0027] Step 2-1-5, accumulate the number of boxes for all grids to obtain the total number of boxes N(s i ) covering the entire image;

[0028] Step 2-1-6, iterate through all scale parameters in the optimal scale parameter combination, repeat steps 2-1-1 to 2-1-5, and obtain the N(s i ) sequence corresponding to different s i ;

[0029] Step 2-1-7, perform least squares linear fitting on log(N(s i )) and log(1 / s i ), and the slope of the fitted line is the fractal dimension D.

[0030] Further, the step 4 specifically comprises:

[0031] Step 4-1, standardize the fractal dimension data and corresponding energy consumption data in the feature database to construct a feature matrix;

[0032] Step 4-2, use the elbow rule to calculate the sum of squared errors within clusters for different K values, and select the K value corresponding to the point where the error squared decreases sharply as the optimal cluster number;

[0033] Step 4-3, the K-means+ algorithm is used to initialize the cluster center, the maximum number of iterations is set, the Euclidean distance is used as the similarity measure, the cluster analysis is carried out on the feature matrix, and the iteration is carried out to the maximum number of iterations, and the fractal dimension mean value, standard deviation and energy consumption average value of each cluster are output; step 4-4, based on the clustering result, the cluster with the highest energy consumption average value is identified as the high energy consumption cluster;

[0034] Step 4-5, based on the fractal dimension mean value of the high energy consumption cluster and the fractal dimension standard deviation of the high energy consumption cluster, the fractal dimension threshold interval of the high energy consumption cluster is calculated.

[0035] Further, in step 4-5, the fractal dimension threshold interval of the high energy consumption cluster is [μ high -2σ high , μ high +2σ high ]; wherein, μ high is the fractal dimension mean value of the high energy consumption cluster, and σ high is the fractal dimension standard deviation of the high energy consumption cluster.

[0036] Further, the adjustment of the running mode of the self-unloading vehicle in step 6 includes:

[0037] Based on the comparison result, when the real-time fractal dimension falls into the threshold interval of the high energy consumption cluster, a running path optimization suggestion or a speed control strategy is generated;

[0038] The running path optimization suggestion or the speed control strategy is sent to the terminal scheduling system to adjust the running parameters of the self-unloading vehicle.

[0039] Further, the terminal self-unloading vehicle running mode optimization method based on the differential box algorithm further includes:

[0040] The running state of the self-unloading vehicle is monitored in real time, and the abnormal fluctuation is detected by continuously analyzing the deviation of the real-time fractal dimension relative to the fractal dimension threshold interval of the high energy consumption cluster;

[0041] When it is detected that the real-time fractal dimension continuously deviates from the fractal dimension threshold interval of the high energy consumption cluster, a warning signal is sent to the monitoring platform;

[0042] Based on the warning signal, an artificial intervention or an automatic adjustment mechanism is triggered.

[0043] The application also provides a terminal self-unloading vehicle running mode optimization system based on the differential box algorithm, which is used to execute the above-mentioned terminal self-unloading vehicle running mode optimization method based on the differential box algorithm, and includes:

[0044] The picture collection module is used for collecting multiple groups of self-unloading vehicle running process pictures and pre-processing;

[0045] Fractal dimension calculation module: for calculating the fractal dimension based on the pre-processed multiple groups of dump truck running process pictures using the differential box algorithm;

[0046] Feature database construction module: for constructing a feature database based on the pre-processed multiple groups of dump truck running process pictures, the feature database including: the fractal dimension corresponding to each of the dump truck running process pictures and the corresponding energy consumption data;

[0047] Threshold interval setting module: for identifying the relationship between the fractal dimension and the energy consumption data using K-means clustering analysis, and setting the fractal dimension threshold interval of the high energy consumption cluster according to the relationship;

[0048] Real-time fractal dimension calculation module: for calculating the real-time fractal dimension according to the real-time collected dump truck running process pictures;

[0049] Dump truck running mode adjustment module: for comparing the real-time fractal dimension with the fractal dimension threshold interval of the high energy consumption cluster, and adjusting the dump truck running mode according to the comparison result.

[0050] The present application has the following technical effects:

[0051] 1. The fractal dimension features of the dump truck running images are extracted by the improved differential box algorithm, the quantitative correlation between the fractal dimension and the energy consumption state is established, and the non-contact intelligent evaluation of the running mode is realized. The running mode optimization is realized through image analysis, the energy consumption and pollutant emission during the vehicle running process are reduced, compared with the complex vehicle position and energy consumption data statistical analysis, the implementation cost is reduced while the system scalability is improved. The real-time fractal dimension calculation and threshold comparison mechanism enables the energy consumption optimization to change from post-analysis to dynamic regulation, realizing the prospective optimization of the running mode.

[0052] 2. The scale parameter combination of the differential box algorithm is adaptively optimized by the genetic algorithm, which significantly improves the stability and differentiation of the fractal dimension calculation, making the energy consumption state recognition more accurate and reliable. The fractal dimension threshold interval set based on K-means clustering analysis can effectively distinguish the energy consumption level, providing a scientific basis for running strategy adjustment.

[0053] 3. Through the fusion of visual feature analysis and intelligent algorithm, a new technical path is provided for port equipment energy consumption management. The non-contact monitoring method reduces the dependence on hardware equipment and improves the system reliability. The dynamic optimization mechanism ensures the continuous effectiveness of energy consumption management, providing key technical support for green port construction. The overall scheme realizes the fine management of energy consumption under the premise of ensuring the operation efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the specific embodiments or the prior art of the present application, the drawings required to be used in the specific embodiments or the prior art description will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0055] Figure 1 is a flow chart of the wharf self-unloading truck running mode optimization method based on the differential box algorithm provided by the embodiment of the present application.

[0056] Figure 2 is a schematic diagram of the wharf self-unloading truck running mode optimization system based on the differential box algorithm provided by the embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely below. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0058] Figure 1 is a flow chart of the wharf self-unloading truck running mode optimization method based on the differential box algorithm provided by the embodiment of the present application. Referring to Figure 1 , the present application provides a wharf self-unloading truck running mode optimization method based on the differential box algorithm, comprising the following steps:

[0059] Step 1, collect multiple groups of self-unloading truck running process pictures and perform preprocessing; the self-unloading truck running process picture is an MxM pixel picture;

[0060] In some embodiments, in step 1, the preprocessing includes: performing gray scale processing, image size standardization and image enhancement processing on the multiple groups of self-unloading truck running process pictures.

[0061] It is worth noting that when collecting multiple groups of self-unloading truck running process pictures, different working scenes such as yard loading and unloading, road driving and other typical working conditions should be covered to ensure the representativeness and diversity of the data. Picture acquisition is usually completed by deploying high-definition cameras at key points on the wharf. These cameras need to have the ability to adapt to complex lighting conditions to avoid image quality degradation due to weather changes or insufficient light. The collected pictures need to be systematically preprocessed to eliminate noise, unify the format and provide clear and consistent input for subsequent fractal dimension calculation.

[0062] The first step of preprocessing is grayscale processing, which converts color images into grayscale images. This process extracts the brightness information of the image, reducing data complexity while preserving key texture features. For example, the outline and motion trajectory of the dump truck in the image are more prominent through grayscale contrast, avoiding the interference of color information on the analysis results. Grayscale enhances the stability of image features, making pictures taken at different times comparable.

[0063] Next is image size standardization, which aims to adjust all pictures to a uniform size, eliminating the problem of inconsistent scales caused by differences in shooting distance or angle. This step can be achieved through interpolation or cropping algorithms, ensuring that each image remains consistent in spatial dimensions, thereby avoiding deviations in fractal dimension calculation due to size variations.

[0064] Image enhancement processing is a key step in preprocessing, focusing on improving the visual quality of the image. Through techniques such as contrast adjustment and histogram equalization, enhancement processing can highlight the details of the dump truck's running path and suppress background interference. For example, in low light conditions, enhancement algorithms can effectively improve image brightness and clarity, making vehicle motion features more obvious. This processing improves the availability of image information, providing a more reliable basis for fractal dimension extraction.

[0065] Step 2: Based on the multiple sets of pre-processed dump truck running process pictures, the improved differential box algorithm is used to calculate the fractal dimension.

[0066] In some embodiments, in step 2, the improved differential box algorithm refers to using a genetic algorithm to obtain the optimal scale parameter combination of the differential box algorithm.

[0067] In the implementation of the dump truck running mode optimization process, the core of step two is to extract fractal dimension features from pre-processed images through the differential box algorithm.

[0068] Step 2 uses a genetic algorithm to determine the best parameter combination of the differential box algorithm. The implementation of the genetic algorithm begins with generating an initial population containing multiple scale parameter combinations, each containing a set of different scale values that need to cover different analysis levels from fine to coarse granularity. For each scale parameter combination in the population, calculate the fractal dimension sequence of multiple sample images, and evaluate the fitness of each parameter combination by analyzing the standard deviation of these sequences and the profile coefficient in different energy consumption intervals. Fitness calculation considers both the stability of fractal dimension calculation and the ability to distinguish different energy consumption states, ensuring that the optimized parameter combination can guarantee the consistency of the results and effectively identify different running modes.

[0069] In some embodiments, the use of a genetic algorithm to obtain the optimal scale parameter combination of the differential box algorithm specifically includes:

[0070] Step 2-1, randomly generate a plurality of scale parameter combinations as an initial population, each scale parameter combination containing a plurality of different scale parameters;

[0071] Step 2-2, for each scale parameter combination in the population, calculate the fractal dimension sequence of a plurality of sample image groups using the differential box algorithm, and calculate the fitness value based on the standard deviation of the fractal dimension and the contour coefficient in different energy consumption data intervals;

[0072] Step 2-3, select the next generation population according to the fitness value and a preset proportion;

[0073] Step 2-4, perform crossover and mutation operations on the selected next generation population to generate new scale parameter combinations;

[0074] Step 2-5, repeat steps 2-2 to 2-4 until the maximum number of iterations is reached, and output the optimal scale parameter combination.

[0075] Specifically, the implementation of the genetic algorithm begins with generating an initial population containing a plurality of scale parameter combinations, each combination containing a set of different scale parameters s1, s2, …, s n These scale parameters need to satisfy the constraint condition: 2≤s i ≤ M / 2. For each scale parameter combination in the population, the fractal dimension sequence of a plurality of sample image groups is calculated, and the fitness value is calculated by the following formula:

[0076] The fitness function is: F= ɑ×(1 / σ) + β×Ss; where F is the fitness value, σ is the standard deviation of the fractal dimension sequence, Ss is the contour coefficient, and ɑ and β are weight coefficients. Based on the fitness value, select the individuals with good performance according to a preset proportion to form the next generation population. Then, the selected individuals perform crossover and mutation operations to generate new scale parameter combinations. This evolution process is repeated until the maximum number of iterations is reached, and the optimal scale parameter combination obtained after optimization is finally output.

[0077] After obtaining the optimal scale parameter combination, the specific calculation process of the differential box algorithm is started.

[0078] In some embodiments, in step 2, the fractal dimension is calculated using an improved differential box algorithm, which specifically includes:

[0079] Step 2-1-1, divide the preprocessed dump truck running process picture into a plurality of grids with a size of s i ×s i , where s i is a scale parameter, and the value range of s i satisfies 2≤s i ≤M / 2;

[0080] Step 2-1-2, for each grid, calculate the maximum and minimum gray scale of the pixels within it;

[0081] Step 2-1-3, based on the total gray scale G of the image, determine the box height k; that is ; wherein ceil() is the ceiling function;

[0082] Step 2-1-4, according to the maximum and minimum gray scale, calculate the number of boxes required to cover each s i ×s i grid; ; n ij is the number of boxes required to cover each s i ×s i grid, max is the maximum gray scale of the pixels within each grid, min is the minimum gray scale of the pixels within each grid, and floor() is the floor function;

[0083] Step 2-1-5, accumulate the number of boxes for all grids to obtain the total number of boxes N(s i ) covering the entire image, that is ;

[0084] Step 2-1-6, traverse all scale parameters in the optimal scale parameter combination, repeat steps 2-1-1 to 2-1-5 to obtain the corresponding N(s i ) sequence for different s i ;

[0085] By traversing all scale parameters in the optimal scale parameter combination and repeating the above grid division and box counting process, a series of total box numbers corresponding to different scale parameters can be obtained, forming the N(s i ) sequence;

[0086] Step 2-1-7, perform least squares linear fitting on log(N(s i )) and log(1 / s i ), and the slope of the fitted straight line is the fractal dimension D.

[0087] The above calculation process significantly improves the representation ability of the fractal dimension feature through the optimized scale parameter combination. This fractal dimension calculation method based on genetic algorithm parameter optimization not only improves the accuracy of feature extraction, but also enhances the adaptability of the algorithm to complex operating environments, providing high-quality feature input for subsequent energy consumption analysis and operation mode optimization.

[0088] Step 3, based on the pre-processed multiple groups of dump truck running process pictures, a feature database is constructed, which includes the fractal dimension corresponding to each dump truck running process picture and the corresponding energy consumption data.

[0089] In the embodiment, the construction of the feature database specifically includes:

[0090] Step 3-1, vehicle running data synchronized with the pre-processed multiple groups of dump truck running process pictures is obtained, which includes but is not limited to instantaneous fuel consumption, cumulative energy consumption or unit mileage energy consumption data of the dump truck.

[0091] Step 3-2, time alignment processing is performed on the vehicle running data to establish the time sequence correspondence between the dump truck running process pictures and the energy consumption data.

[0092] Step 3-3, the image identifier of each dump truck running process picture, the corresponding fractal dimension and the corresponding energy consumption data are taken as a data record.

[0093] Step 4, the relationship between the fractal dimension and the energy consumption data is identified by K-means clustering analysis, and the fractal dimension threshold interval of the high energy consumption cluster is set according to the relationship.

[0094] Step 4 establishes the quantitative relationship between the fractal dimension and the energy consumption data through clustering analysis, and sets a scientific fractal dimension threshold interval for the high energy consumption cluster.

[0095] Further, step 4 specifically includes:

[0096] Step 4-1, the fractal dimension data and the corresponding energy consumption data in the feature database are standardized to construct a feature matrix.

[0097] Optionally, the Z-score standardization method is used to process the fractal dimension data and the corresponding energy consumption data in the feature database, eliminating the dimensional differences between the original features. The standardized data is organized in the form of a matrix as the input of K-means clustering.

[0098] Step 4-2, the elbow rule is used to calculate the within-cluster sum of squares under different K values, and the K value corresponding to the point where the error sum of squares decreases sharply is selected as the optimal cluster number; by calculating the within-cluster sum of squares corresponding to multiple K values, the inflection point position of the error decrease curve is observed, and the point where the decrease rate changes significantly is selected as the optimal cluster number. This avoids the randomness of subjective selection of the number of clusters and ensures the objectivity of the clustering results.

[0099] Step 4-3, using K-means+ algorithm to initialize cluster centers, setting the maximum number of iterations, using Euclidean distance as similarity measure, clustering analysis on the feature matrix, iterating to the maximum number of iterations, outputting the fractal dimension mean, standard deviation and energy consumption average of each cluster.

[0100] Step 4-4, based on the clustering results, the cluster with the highest energy consumption average is identified as the high energy consumption cluster.

[0101] Step 4-5, based on the fractal dimension mean of the high energy consumption cluster and the fractal dimension standard deviation of the high energy consumption cluster, the fractal dimension threshold interval of the high energy consumption cluster is calculated.

[0102] In some embodiments, the fractal dimension threshold interval of the high energy consumption cluster is [μ high -2σ high , μ high +2σ high ]; wherein μ high is the fractal dimension mean of the high energy consumption cluster, and σ high is the fractal dimension standard deviation of the high energy consumption cluster. The threshold interval of the fractal dimension mean and the standard deviation of the high energy consumption cluster is calculated respectively. The fractal dimension threshold interval of the high energy consumption cluster is based on the range from the fractal dimension mean minus twice the standard deviation to the fractal dimension mean plus twice the standard deviation. This interval setting ensures that it covers most of the data points in the cluster, providing a reliable classification boundary. The establishment of the threshold interval enables the system to evaluate the running state of the dump truck in real time, providing a basis for subsequent optimization decisions.

[0103] Step 5, according to the real-time collected dump truck running process pictures, the real-time fractal dimension is calculated.

[0104] Step 6, comparing the real-time fractal dimension with the fractal dimension threshold interval of the high energy consumption cluster, and adjusting the running mode of the dump truck according to the comparison result.

[0105] In some embodiments, the step 6 of adjusting the running mode of the dump truck comprises:

[0106] based on the comparison result, when the real-time fractal dimension falls into the fractal dimension threshold interval of the high energy consumption cluster, generating a running path optimization suggestion or a speed control strategy;

[0107] sending the running path optimization suggestion or the speed control strategy to the terminal dispatching system to adjust the running parameters of the dump truck.

[0108] Specifically, after the real-time fractal dimension calculation is completed, the fractal dimension threshold interval of the corresponding high energy consumption cluster is called. The comparison logic is based on a simple interval judgment: if the real-time fractal dimension falls into the fractal dimension threshold interval of the high energy consumption cluster, the optimization mechanism is triggered.

[0109] Optionally, when the real-time fractal dimension reaches the high energy consumption interval, the suggestion generation module will be immediately started. The running path optimization suggestion mainly focuses on the rationality of the driving route, and recommends a more direct driving route or a more reasonable operation sequence by analyzing the efficient path mode in the historical data. The speed control strategy focuses on the adjustment of the running rhythm, and suggests to use a more stable acceleration curve and a more economical cruise speed. The running state after adjustment is continuously monitored, and the optimization effect is verified by continuous fractal dimension calculation. If the adjusted fractal dimension is still in the high energy consumption interval, further optimization program will be started.

[0110] In some embodiments, the wharf self-unloading truck running mode optimization method based on the differential box algorithm further comprises:

[0111] Real-time monitoring of the running state of the self-unloading truck, detecting abnormal fluctuations by continuously analyzing the deviation of the real-time fractal dimension from the fractal dimension threshold interval of the high energy consumption cluster;

[0112] When the real-time fractal dimension continuously deviates from the fractal dimension threshold interval of the high energy consumption cluster, a warning signal is sent to the monitoring platform;

[0113] Based on the warning signal, triggering manual intervention or automatic adjustment mechanism.

[0114] When monitoring the running state of the self-unloading truck, the fractal dimension of the self-unloading truck running process picture is calculated in real time, and it is compared with the pre-set fractal dimension threshold interval of the high energy consumption cluster. The monitoring process focuses on analyzing the continuous change trend of the fractal dimension, and when the real-time value continuously deviates from the normal range (long time in the fractal dimension threshold interval of the high energy consumption cluster), it is marked as an abnormal state. The abnormal degree is evaluated by calculating the deviation amplitude and duration of the fractal dimension from the threshold interval. For example, if the fractal dimension continuously exceeds the threshold interval for multiple periods, a warning signal will be generated. The warning signal is sent to the central monitoring platform, and the abnormal details including vehicle number, deviation time and suggested measures are displayed in a visual form. Based on the warning signal, a two-level response mechanism is triggered: for slight deviation, an automatic adjustment program is started, such as fine-tuning the running speed or path planning; for serious or continuous abnormality, the operator is notified to intervene for manual intervention to ensure timely correction of the running mode. The whole monitoring process runs in a closed loop, and the self-unloading truck energy consumption performance is optimized through continuous feedback.

[0115] The present application is aimed at the typical links of loading, unloading and horizontal transportation of the self-unloading truck at the bulk cargo terminal, and the running mode features are accurately captured through continuous image acquisition; the core is to calculate the fractal dimension of the image by using the improved differential box algorithm, and to establish a quantitative correlation model of the fractal dimension and the energy consumption data based on K-means clustering analysis, so as to realize the non-contact intelligent evaluation of the running state. By comparing the fractal dimension with the preset threshold interval in real time, the speed control or path optimization strategy is dynamically generated and sent to the scheduling system, so as to realize the change from post-event statistics to in-event intervention, significantly improve the energy consumption efficiency, and reduce the operation cost and emissions.

[0116] As shown in Figure 2 The present application also provides a terminal self-unloading truck running mode optimization system based on the differential box algorithm, which is used to execute the terminal self-unloading truck running mode optimization method based on the differential box algorithm, and includes:

[0117] An image collection module is used to collect multiple groups of self-unloading truck running process pictures and perform preprocessing;

[0118] A fractal dimension calculation module is used to calculate the fractal dimension based on the multiple groups of preprocessed self-unloading truck running process pictures by using the differential box algorithm;

[0119] A feature database construction module is used to construct a feature database based on the multiple groups of preprocessed self-unloading truck running process pictures, and the feature database includes the fractal dimension corresponding to each self-unloading truck running process picture and the corresponding energy consumption data;

[0120] A threshold interval setting module is used to identify the relationship between the fractal dimension and the energy consumption data by using K-means clustering analysis, and to set the fractal dimension threshold interval of the high energy consumption cluster according to the relationship;

[0121] A real-time fractal dimension calculation module is used to calculate the real-time fractal dimension according to the real-time collected self-unloading truck running process pictures;

[0122] A self-unloading truck running mode adjustment module is used to compare the real-time fractal dimension with the fractal dimension threshold interval of the high energy consumption cluster, and to adjust the self-unloading truck running mode according to the comparison result.

[0123] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the embodiments of the present application.

Claims

1. A method for optimizing the operation mode of terminal dump trucks based on the differential box algorithm, characterized in that, Includes the following steps: Step 1: Collect multiple sets of images of the dump truck in operation and preprocess them; the images of the dump truck in operation are M×M pixel images; Step 2: Based on the preprocessed images of multiple sets of dump truck operation processes, the fractal dimension is calculated using an improved difference box algorithm; Step 3: Based on the preprocessed multiple sets of dump truck operation process images, construct a feature database, which includes: the fractal dimension and corresponding energy consumption data for each dump truck operation process image; Step 4: Use K-means clustering analysis to identify the relationship between fractal dimension and energy consumption data, and set the fractal dimension threshold range for high energy consumption clusters based on the relationship. Step 5: Calculate the real-time fractal dimension based on the real-time images of the dump truck's operation. Step 6: Compare the real-time fractal dimension with the fractal dimension threshold range of the high-energy-consumption cluster, and adjust the dump truck operation mode according to the comparison result; In step 2, the improved differential box algorithm refers to using a genetic algorithm to obtain the optimal combination of scale parameters for the differential box algorithm. The process of using a genetic algorithm to obtain the optimal combination of scale parameters for the difference box algorithm specifically includes: Step 2-1: Randomly generate multiple scale parameter combinations as the initial population. Each scale parameter combination contains several different scale parameters. Step 2-2: For each combination of scale parameters in the population, the difference box algorithm is used to calculate the fractal dimension sequence of multiple sets of sample images, and the fitness value is calculated based on the standard deviation of the fractal dimension and the contour coefficient in different energy consumption data intervals. Steps 2-3: Select the next generation population based on fitness values ​​and preset ratios; Steps 2-4: Perform crossover and mutation operations on the selected next generation population to generate new combinations of scale parameters; Step 2-5: Repeat steps 2-2 to 2-4 until the maximum number of iterations is reached, and output the optimal combination of scale parameters; Step 4 specifically includes: Step 4-1: Standardize the fractal dimension data and corresponding energy consumption data in the feature database to construct a feature matrix; Step 4-2: Use the elbow rule to calculate the sum of squared errors within clusters for different K values, and select the K value corresponding to the point where the rate of decrease of the sum of squared errors changes abruptly as the optimal number of clusters; Step 4-3: Initialize cluster centers using the K-means algorithm, set the maximum number of iterations, use Euclidean distance as a similarity measure, perform cluster analysis on the feature matrix, iterate to the maximum number of iterations, and output the mean fractal dimension, standard deviation, and average energy consumption of each cluster. Step 4-4: Based on the clustering results, identify the cluster with the highest average energy consumption as a high-energy-consuming cluster; Steps 4-5: Based on the mean fractal dimension and standard deviation of the fractal dimension of the high-energy-consuming cluster, calculate the fractal dimension threshold range of the high-energy-consuming cluster.

2. The method for optimizing the operation mode of terminal dump trucks based on the differential box algorithm according to claim 1, characterized in that, In step 1, the preprocessing includes: performing grayscale processing, image size standardization, and image enhancement processing on the multiple sets of dump truck operation process images.

3. The method for optimizing the operation mode of terminal dump trucks based on the differential box algorithm according to claim 1, characterized in that, In step 2, the fractal dimension is calculated using an improved difference box algorithm, specifically including: Step 2-1-1: Divide the pre-processed dump truck operation images into multiple images of size s. i ×s i A grid of pixels, where s i Let be the scale parameter, and s i The range of values ​​for s satisfies 2 ≤ s i ≤M / 2; Step 2-1-2: For each grid cell, calculate the maximum and minimum grayscale values ​​of the pixels within it. Step 2-1-3: Determine the box height k based on the total gray level G of the image; Step 2-1-4: Calculate the number of boxes required to cover each grid based on the maximum and minimum grayscale values; Step 2-1-5: Sum the number of boxes in all grids to obtain the total number of boxes N(s) covering the entire image. i ); Step 2-1-6: Iterate through all scale parameters in the optimal scale parameter combination, repeating steps 2-1-1 to 2-1-5 to obtain different s i The corresponding N(s) i )sequence; Step 2-1-7, calculate log(N(s) i )) and log(1 / s i Perform a least-squares linear fit, and the slope of the fitted line is the fractal dimension D.

4. The method for optimizing the operation mode of terminal dump trucks based on the differential box algorithm according to claim 1, characterized in that, In steps 4-5, the fractal dimension threshold range for high-energy-consuming clusters is [μ]. high -2σ high μ high +2σ high ]; where μ high Let σ be the mean fractal dimension of the high-energy-consuming cluster. high is the standard deviation of the fractal dimension of the high-energy-consuming cluster.

5. The method for optimizing the operation mode of terminal dump trucks based on the differential box algorithm according to claim 1, characterized in that, Step 6, adjusting the dump truck's operating mode, includes: Based on the comparison results, when the real-time fractal dimension falls into the threshold range of the high-energy-consuming cluster, a running path optimization suggestion or speed control strategy is generated. The proposed route optimization or speed control strategy is sent to the terminal scheduling system to adjust the operating parameters of the dump truck.

6. The method for optimizing the operation mode of terminal dump trucks based on the differential box algorithm according to claim 1, characterized in that, The method for optimizing the operation mode of terminal dump trucks based on the differential box algorithm also includes: The system monitors the operating status of the dump truck in real time and detects abnormal fluctuations by continuously analyzing the deviation of the real-time fractal dimension from the fractal dimension threshold range. When the real-time fractal dimension is detected to continuously deviate from the fractal dimension threshold range, an early warning signal is sent to the monitoring platform; Based on the aforementioned warning signal, a manual intervention or automatic adjustment mechanism is triggered.

7. A terminal dump truck operation mode optimization system based on differential box algorithm, used to execute the terminal dump truck operation mode optimization method based on differential box algorithm as described in any one of claims 1-6, characterized in that, The system includes: Image collection module: Used to collect multiple sets of images of the dump truck in operation and perform preprocessing; Fractal dimension calculation module: Used to calculate the fractal dimension based on multiple pre-processed images of dump truck operation, using the difference box algorithm. Feature database construction module: used to construct a feature database based on multiple pre-processed images of dump truck operation process, the feature database including: the fractal dimension and corresponding energy consumption data of each image of dump truck operation process; Threshold interval setting module: used to identify the relationship between fractal dimension and energy consumption data using K-means clustering analysis, and to set the fractal dimension threshold interval for high energy consumption clusters based on the relationship; Real-time fractal dimension calculation module: used to calculate the real-time fractal dimension based on real-time images of the dump truck's operation process; The dump truck operation mode adjustment module is used to compare the real-time fractal dimension with the fractal dimension threshold range of the high-energy-consumption cluster, and adjust the dump truck operation mode according to the comparison result.

Images