Cleaning and sweeping vehicle operation optimization method and system based on intelligent identification, medium and product

Abstract

The invention provides a cleaning and sweeping vehicle operation optimization method and system based on intelligent recognition, a medium and a product, and relates to the technical field of intelligent control. The method comprises the following steps: acquiring pavement image information by using a plurality of garbage monitors; then, using a garbage classification model subjected to deep learning training to determine garbage category information according to the road surface image; then, geometric dimension data and garbage distribution density information are obtained in combination with the road surface image information, and garbage volume information is calculated; determining weight coefficient information according to a garbage category weight model; obtaining pavement garbage equivalent information by integrating the volume, the weight coefficient and the distribution density information; and working parameters of the rolling brush assembly motor and the high-pressure water pump are adjusted according to a set strategy to optimize the operation of the cleaning and sweeping vehicle. By implementing the method, the working parameters of the rolling brush assembly motor and the high-pressure water pump can be intelligently and efficiently adjusted, so that the cleaning and sweeping vehicle operation is more targeted, and the cleaning efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent control technology, and particularly relates to a method, system, medium and product for optimizing the operation of a washing and sweeping vehicle based on intelligent recognition. Background Art

[0002] In the field of urban environmental sanitation maintenance, keeping the roads clean is crucial for enhancing the urban image and ensuring the quality of life of residents. With the continuous expansion of the urban scale and the continuous increase in population, the task volume of road cleaning is becoming increasingly heavy, and the requirements for cleaning efficiency and effect are also getting higher and higher. In order to achieve efficient road cleaning operations, it is necessary to accurately grasp the situation of road garbage so as to reasonably arrange cleaning resources.

[0003] Currently, most sweeping vehicles adopt relatively traditional cleaning methods. Operators, relying on their own experience, preset the working parameters of the roller brush assembly motor and the high-pressure water pump according to the general situation of the road, such as the prosperity degree of the road, the surrounding environment, etc. For example, in a bustling commercial street, due to the large flow of people and relatively more garbage generated, the operator will increase the roller brush speed and the pressure of the high-pressure water pump; while in a relatively remote and less garbage street, appropriately reduce the roller brush speed and the water pump pressure. At the same time, for the detection of garbage, it mainly relies on manual observation. When the operator finds garbage on the road surface, the cleaning operation begins.

[0004] However, the urban road environment is complex and changeable, and there are great differences in the types, quantities and distribution of garbage generated in different regions and at different times. For example, after holding a large-scale event, there may be a large amount of different types of garbage on the road, including both light garbage such as food packaging and relatively large garbage such as discarded promotional materials. It is very difficult for the working parameters set by manual experience to adapt to this complex and changeable garbage situation, resulting in low cleaning efficiency and incomplete cleaning of some garbage. When encountering the situation of suddenly increased garbage or relatively scattered distribution, it is impossible to detect and adjust the cleaning strategy in time, seriously affecting the cleaning effect. Therefore, the traditional method is difficult to accurately obtain the relevant information of road garbage and intelligently and efficiently adjust the working parameters of the sweeping vehicle. Summary of the Invention

[0005] The present application provides a method, system, medium and product for optimizing the operation of a washing and sweeping vehicle based on intelligent recognition, which is used to realize the intelligence, high efficiency and precision of the operation of the washing and sweeping vehicle.

[0006] In a first aspect, the present application provides an operation optimization method for a washing and sweeping vehicle based on intelligent recognition, which is applied to an operation optimization system. The method includes: obtaining road surface image information through multiple garbage monitors; combining the road surface image information, and determining garbage category information through a garbage classification model, which is obtained by deep learning training in advance with multiple garbage image data labeled with category information; combining the road surface image information to determine geometric dimension data and garbage distribution density information corresponding to the garbage category information, and determining garbage volume information through the geometric dimension data; combining the garbage category information, and determining multiple weight coefficient information according to a garbage category weight model, which is obtained by deep learning training in advance with multiple garbage image data labeled with weight coefficients; combining the garbage volume information, the weight coefficient information and the distribution density information to determine road surface garbage equivalent information; and adjusting the working parameters of the roller brush assembly motor and the high-pressure water pump according to the road surface garbage equivalent according to a set adjustment strategy.

[0007] By adopting the above technical solution, a comprehensive road surface image information can be obtained by using multiple garbage monitors, and the garbage classification model trained by deep learning can accurately determine the garbage category information. The geometric dimensions and garbage distribution density can be obtained by combining the image information, and then the garbage volume information can be calculated. Then, the weight coefficient information is determined according to the garbage category weight model. These information comprehensively determine the road surface garbage equivalent information, providing accurate data support for subsequent operations. In this way, the working parameters of the roller brush assembly motor and the high-pressure water pump can be accurately adjusted, making the operation of the washing and sweeping vehicle more targeted, improving the cleaning efficiency, reducing resource waste, and enhancing the operation effect.

[0008] In combination with some embodiments of the first aspect, in some embodiments, after the step of combining the garbage volume information, the weight coefficient information and the distribution density information to determine the road surface garbage equivalent information, it further includes: determining weather condition information and road condition information through multiple sensors; combining the weather condition information and the road condition information, and correcting the road surface garbage equivalent information according to a preset garbage equivalent correction coefficient mapping database. The correction at least includes that when the sensor detects that it is rainy currently and the road surface humidity is greater than a preset humidity threshold, it is determined that there is rain causing the garbage weight to increase and the garbage to stick to the road surface, and then a corresponding correction coefficient is obtained from the garbage equivalent correction coefficient mapping database to increase the road surface garbage equivalent information.

[0009] By adopting the above technical solution, the influence of weather and road conditions on garbage cleaning is considered, making the road surface garbage equivalent information more in line with the actual situation, so as to ensure that the washing and sweeping vehicle adjusts the operation intensity according to the actual situation, avoiding incomplete cleaning caused by weather factors, and improving the scientificity and effectiveness of the operation.

[0010] In some embodiments in combination with some embodiments of the first aspect, after the step of determining the road surface garbage equivalent information by combining the garbage volume information, the weight coefficient information, and the distribution density information, the following steps are further included: If the road surface garbage equivalent information is greater than a preset equivalent threshold, a warning message is sent to the user terminal, and at the same time, the cleaning equipment is started according to a preset additional cleaning plan to enhance the cleaning measures.

[0011] By adopting the above technical solutions, the early warning and active adjustment mechanism can avoid a large accumulation of garbage and ensure the cleanliness of the road surface. At the same time, through timely strengthening of the cleaning equipment, more in-depth cleaning work can be carried out when the amount of garbage is large, effectively preventing the adverse effects of garbage on the environment and traffic, and ensuring that the cleaning work always meets the environmental quality requirements.

[0012] In some embodiments in combination with some embodiments of the first aspect, after the step of adjusting the working parameters of the rotary brush assembly motor and the high-pressure water pump according to the road surface garbage equivalent according to a set adjustment strategy, the following steps are further included: obtaining the expected garbage cleaning effect within a set time, where the expected garbage cleaning effect refers to the garbage equivalent that the sweeper is expected to clean within the set time after being adjusted according to the set adjustment strategy; obtaining the actual garbage cleaning effect within the set time; if the actual garbage cleaning effect is less than the expected garbage cleaning effect, the actual garbage cleaning effect and the expected garbage cleaning effect are used to determine weight adjustment information through a weight adjustment model, and the weight adjustment model is obtained through deep learning training in advance based on multiple actual garbage cleaning effects and expected garbage cleaning effects with marked weight adjustment information; adjusting the garbage category weight model according to the weight adjustment information.

[0013] By adopting the above technical solutions, the expected and actual garbage cleaning effects within a set time are obtained. When the actual effect is lower than the expected effect, the weight adjustment model is used to adjust the garbage category weight model. This helps to optimize the garbage category weight model, enabling it to be dynamically adjusted according to the actual cleaning situation, making the subsequent adjustment of working parameters according to the road surface garbage equivalent more reasonable, improving the allocation efficiency of cleaning resources, avoiding cleaning effect deviations, realizing the continuous optimization of the cleaning work, and ensuring the long-term stability of the cleaning effect.

[0014] In some embodiments in combination with some embodiments of the first aspect, after the step of determining the garbage category information through a garbage classification model, the following steps are further included: performing spatio-temporal clustering analysis on the garbage category information to obtain the garbage category distribution characteristics in different regions and different time periods; establishing a garbage generation prediction model according to the garbage category distribution characteristics, and the prediction model is used to predict the garbage category and quantity generated in a set region within a set time period; when the prediction result shows that the garbage generation amount in a certain region in a future time period will exceed a preset threshold, the cleaning frequency and intensity of the region are adjusted in advance.

[0015] By adopting the above technical solutions, spatio-temporal clustering analysis is carried out on the garbage category information, and a garbage generation prediction model is established. This model can predict the garbage category and quantity in a set area and set time period. When the prediction result shows that the garbage quantity exceeds the threshold, the cleaning frequency and intensity are adjusted in advance. This can plan the cleaning work in advance, reasonably allocate resources, avoid the phenomenon of garbage accumulation, reduce the impact of garbage accumulation on the environment and public health, and at the same time improve the planning and foresight of the cleaning work.

[0016] Combined with some embodiments of the first aspect, in some embodiments, after the step of determining the garbage category information through the garbage classification model in combination with the road surface image information, the following steps are further included: collecting road surface temperature data and road surface material information; when it is detected that the road surface temperature data exceeds the preset temperature threshold, obtaining road surface thermal imaging data; identifying chewing gum residues on the road surface based on the road surface thermal imaging data, and recording the position information, area information and adhesion degree information of the chewing gum residues into a set garbage database; selecting the optimal cleaning agent formula from the cleaning agent formula database according to the road surface temperature data, the road surface material information and the chewing gum residue information, and controlling and adjusting the equipment to automatically adjust the ratio of the cleaning agent to water to obtain a special cleaning solution; real-time monitoring the change of the road surface temperature through a thermal sensor, and when the temperature reaches the set temperature range suitable for removing chewing gum residues, controlling the washing and sweeping vehicle to use the special cleaning solution for a special cleaning procedure.

[0017] By adopting the above technical solutions, it is possible to accurately handle stubborn chewing gum residues according to the road surface conditions. By selecting a special cleaning solution and cleaning at a suitable temperature, the cleaning efficiency of special garbage is improved, and the impact of chewing gum residues on the road surface beauty and cleanliness is avoided.

[0018] Combined with some embodiments of the first aspect, in some embodiments, after the step of determining multiple weight coefficient information according to the garbage category weight model, the following steps are further included: collecting road surface odor data through an odor sensor array; judging whether there is pet feces or rotten organic matter in the target area based on the road surface odor data; when a set odor is detected, starting an ultraviolet spectroscopy analyzer for substance composition analysis, and selecting a special bio-enzyme degrading agent according to the substance composition analysis result; controlling a micro-injection system to release the bio-enzyme degrading agent to the target area; real-time monitoring the degree of chemical reaction during the degradation process, and dynamically adjusting the spraying amount of the enzyme preparation; after the degradation treatment is completed, performing local air purification through an ozone generator to eliminate the residual odor.

[0019] By adopting the above technical solutions, it is possible to accurately handle special garbage, effectively degrade pet feces or rotten organic matter, eliminate odors, purify the local air, improve the environmental quality, and at the same time reduce the adverse effects of special garbage residues on the environment and pedestrians.

[0020] In a second aspect, the present application provides an operation optimization system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, and the computer program code includes computer instructions. The one or more processors call the computer instructions to cause the operation optimization system to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0021] In a third aspect, the present application provides a computer-readable storage medium, including instructions, when the instructions run on an operation optimization system, causing the operation optimization system to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0022] In a fourth aspect, the present application provides a computer program product, when the computer program product runs on an operation optimization system, causing the operation optimization system to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. Since the technical means of obtaining road surface image information based on multiple garbage monitors, determining garbage category information using a garbage classification model trained by deep learning, combining the image information to obtain geometric dimensions, garbage distribution density, garbage volume information, and determining weight coefficient information using a garbage category weight model, and then comprehensively determining the road surface garbage equivalent information are adopted, the technical problem of inaccurate adjustment of cleaning parameters in the existing sweeper operation is effectively solved. Furthermore, the technical effects of strong pertinence, high cleaning efficiency, less resource waste, and good operation effect of the sweeper cleaning operation are achieved.

[0024] 2. Since the technical means of, after determining the road surface garbage equivalent information, combining multiple sensors to obtain weather and road condition information, and when it is rainy and the road surface humidity exceeds the threshold, obtaining a correction coefficient from a garbage equivalent correction coefficient mapping database to correct the garbage equivalent information are adopted, the technical problem that the influence of weather and road conditions on the cleaning effect is not considered in the existing sweeper operation is effectively solved. Furthermore, the technical effect of scientifically adjusting the intensity of the sweeper operation according to the actual situation and avoiding incomplete cleaning is achieved.

[0025] 3. By adopting the technical means of collecting road surface temperature and material information, obtaining thermal imaging data to identify chewing gum residues when the road surface temperature exceeds the preset threshold, selecting the optimal cleaning agent formula from the cleaning agent formula database, controlling and adjusting the equipment to prepare a special cleaning solution, and carrying out special cleaning at an appropriate temperature, the technical problem that it is difficult to effectively remove special garbage such as chewing gum in the operation of existing washing and sweeping vehicles is effectively solved. Furthermore, efficient cleaning of special garbage is achieved, and the technical effect of avoiding the influence of chewing gum residues on the road surface beauty and cleanliness is realized. Brief Description of the Drawings

[0026] Figure 1 is a schematic flowchart of an optimization method for the operation of a washing and sweeping vehicle based on intelligent recognition in an embodiment of the present application; Figure 2 is another schematic flowchart of an optimization method for the operation of a washing and sweeping vehicle based on intelligent recognition in an embodiment of the present application; Figure 3 is a schematic structural diagram of an entity device of an operation optimization system in an embodiment of the present application. Detailed Embodiments

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular forms "a", "an", "the", "above-mentioned", "the foregoing", "this" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to and includes any or all possible combinations of one or more of the listed items.

[0028] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as implying or suggesting relative importance or implicitly indicating the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0029] For ease of understanding, the method provided in this embodiment is described in terms of a process below. Please refer to Figure 1 , which is a schematic flowchart of an optimization method for the operation of a washing and sweeping vehicle based on intelligent recognition in an embodiment of the present application.

[0030] S101. Combine the road surface image information and determine the garbage category information through a garbage classification model, which is obtained by deep learning training in advance with multiple garbage image data labeled with category information; The operation optimization system controls multiple garbage monitors mounted on the washing and sweeping vehicle, such as high-resolution cameras, infrared sensors, etc., to comprehensively capture and collect data on the road surface from different angles and at different times. These garbage monitors are distributed at the front of the washing and sweeping vehicle, both sides of the vehicle body, and the rear of the vehicle, etc., and can achieve 360-degree dead-angle-free road surface monitoring. For the large amount of image data collected, the operation optimization system uses a garbage classification model trained by deep learning to process it. This garbage classification model is constructed based on the convolutional neural network (CNN) architecture, such as an improved ResNet model. In the training stage, the system collects a large amount of garbage image data with labeled category information, covering various common garbage types, such as paper, plastic, metal, fruit peels, fallen leaves, etc. These data are randomly divided into a training set, a validation set, and a test set. During the training process, the training set is input into the ResNet model, and the model automatically extracts the features of the garbage images, such as shape, color, texture, etc. After processing through multiple convolutional layers, pooling layers, and fully connected layers, the model outputs the prediction results of the garbage categories. The model training uses the cross-entropy loss function and continuously adjusts the parameters of the model through the backpropagation algorithm to minimize the difference between the prediction results and the labeled true categories. After multiple rounds of iterative training, when the accuracy of the model on the validation set reaches a high level, the training stops, and the model is tested on the test set to evaluate the generalization ability of the model.

[0031] In practical applications, the preprocessed road surface images are input into the trained garbage classification model, and the model quickly outputs the garbage category information. To further improve the accuracy of garbage classification, the system also introduces transfer learning technology. When encountering new garbage types or special scenarios, a small amount of new data is collected to fine-tune the pre-trained model. For example, when a new type of degradable garbage appears in some areas, the system collects the image data of this type of garbage and performs fine-tuning training on the original model so that the model can accurately identify the new type of garbage.

[0032] S102. Determine the geometric dimension data and garbage distribution density information corresponding to the garbage category information in combination with the road surface image information, and determine the garbage volume information through the geometric dimension data; After the operation optimization system obtains the road surface image information that has been sorted by garbage classification, it uses image measurement technology to determine the geometric dimension data of the garbage. The system adopts a pixel-based measurement method. First, it calibrates the image. According to information such as the parameters of the camera and the distance between the washing and sweeping vehicle and the road surface, it establishes a mapping relationship between image pixels and actual dimensions. For garbage with regular shapes, such as rectangular paper, square plastic boxes, etc., the operation optimization system detects the edge contour of the garbage through an edge detection algorithm, and then calculates its length, width and other dimension information according to the pixel coordinates of the contour. For garbage with irregular shapes, such as fallen leaves, fruit peels, etc., the system adopts a polygon approximation algorithm to approximate the irregular shape as a polygon, and determines the approximate size of the garbage by calculating the vertex coordinates of the polygon. When determining the garbage distribution density information, the operation optimization system divides the road surface image into multiple small areas. For example, a cell is 10cm×10cm. It counts the number of garbage in each cell, and combines with the area of the cell to calculate the number of garbage per unit area, that is, the garbage distribution density. To improve the accuracy of density calculation, the system also considers the overlapping situation of garbage. When it detects that there is overlapping garbage, through stereo vision technology, it uses the images taken by multiple garbage monitors from different angles for three-dimensional reconstruction to distinguish the overlapping garbage and accurately count the number of garbage. It calculates the garbage volume information according to the determined geometric dimension data. For garbage with regular shapes, such as objects in the shape of a cuboid, according to its length, width and height dimensions, it directly uses the volume calculation formula to calculate the volume. For garbage with irregular shapes, it adopts an approximate calculation method, such as approximating it as a combination of basic geometric bodies such as an ellipsoid, a cylinder, etc., and calculates the volume according to the dimensions of the approximate geometric bodies.

[0033] S103. Combine the garbage category information and determine multiple weight coefficient information according to the garbage category weight model, which is obtained through deep learning training in advance based on multiple garbage image data with marked weight coefficients; After the operation optimization system obtains the garbage category information, it calls the pre-trained garbage category weight model to determine multiple weight coefficient information. The garbage category weight model is constructed based on deep learning technology, specifically using a multi-layer perceptron (MLP) network structure. In the training stage, the system collects a large amount of garbage image data with marked weight coefficients. These data cover images of various types of garbage in different scenarios, and each image sample is marked with the corresponding weight coefficient, which reflects the comprehensive influence degree of different garbage categories on aspects such as the operation difficulty of the washing and sweeping vehicle, resource consumption, and environmental impact.

[0034] When training the garbage category weight model, the labeled data is divided into a training set, a validation set, and a test set. The training set is used for the parameter learning of the model, the validation set is used to adjust the hyperparameters of the model to prevent overfitting, and the test set is used to evaluate the generalization performance of the model. During the training process, the garbage image data of the training set and its corresponding weight coefficient annotations are input into the MLP network. The MLP network extracts features and performs non-linear transformations on the input data through multiple hidden layers. The neurons in the hidden layer use activation functions to increase the non-linear expression ability of the model. The output layer of the network outputs the predicted values of the weight coefficients. The model training uses the mean squared error (MSE) loss function and adjusts the weight and bias parameters in the network through the backpropagation algorithm to minimize the error between the predicted weight coefficients and the labeled true weight coefficients. After multiple rounds of iterative training, when the loss value of the model on the validation set tends to be stable and meets certain performance indicators, the model training is considered complete and tested on the test set to ensure that the model has good generalization ability.

[0035] In practical applications, the job optimization system inputs the garbage category information obtained from the garbage classification model into the trained garbage category weight model. The model quickly outputs the weight coefficient information corresponding to various types of garbage. To make the weight coefficients more in line with the actual operation situation, the system introduces a dynamic weight adjustment mechanism. Considering the differences in garbage treatment costs, environmental protection requirements, and sweeping vehicle equipment in different regions, the system collects data related to garbage treatment in different regions, including information such as garbage treatment costs, equipment wear conditions, and environmental protection standards. Based on this data, the weight coefficients are dynamically adjusted. For example, in regions with higher environmental protection requirements, for garbage with greater pollution (such as chemical garbage), the weight coefficient is appropriately increased to ensure that more resources are invested when the sweeping vehicle processes this type of garbage to ensure environmental cleanliness; in areas where the equipment is more severely aged, for garbage that is difficult to clean (such as garbage with high viscosity), the weight coefficient is increased to prompt the sweeping vehicle to adopt more appropriate operation parameters when cleaning this type of garbage and reduce equipment wear.

[0036] S104. Determine the road surface garbage equivalent information by combining the garbage volume information, the weight coefficient information, and the distribution density information; After obtaining the garbage volume information, the weight coefficient information, and the distribution density information, the job optimization system determines the road surface garbage equivalent information through a specific algorithm. The road surface garbage equivalent information is a comprehensive measurement index used to reflect the overall impact degree of road surface garbage on the operation of the sweeping vehicle, so as to provide an accurate basis for subsequent adjustment of the sweeping vehicle operation parameters.

[0037] First, the system normalizes the garbage volume information, weight coefficient information, and distribution density information. Since the dimensions and value ranges of these three types of information may be different, in order to make them comparable in calculations, a normalization method is used to map them to the [0, 1] interval. For the garbage volume information, by dividing the volume value of each piece of garbage by the maximum value among all garbage volumes, it is normalized to the [0, 1] range. For the weight coefficient information, since it is itself a relative importance index determined according to the garbage category and already has a certain relative meaning, but for unified calculation with other information, it is also appropriately normalized, for example, by linear transformation to map it to the [0, 1] interval. For the garbage distribution density information, the number of garbage pieces per unit area is divided by a set maximum garbage distribution density value to achieve normalization.

[0038] Then, the job optimization system calculates the road surface garbage equivalent information by means of weighted summation. Let the garbage volume information be V i (i represents different garbage individuals or categories), the weight coefficient information be W i , the distribution density information be D i , and the corresponding information after normalization be V i ′ , W i ′ , D i ′ , and the calculation formula for the road surface garbage equivalent information E is: E = ∑ i aV i ′ W i ′ D i ′ , where a is a balance coefficient, which is determined through a large number of experiments and statistical analysis of actual operation data, and is used to balance the relative importance of volume, weight, and distribution density in calculating the garbage equivalent. This coefficient can be dynamically adjusted according to different road types, operation scenarios, and other factors. For example, on the main roads with heavy traffic, the garbage distribution density has a greater impact on the operation, and the weight related to the distribution density in a is appropriately increased; in some areas with complex garbage types and large differences in the treatment difficulty of different types of garbage, the influence of the weight coefficient in the calculation is increased, and the parameters related to the weight coefficient in a are adjusted.

[0039] To more accurately determine the road surface garbage equivalent information, the operation optimization system also considers the impact of the degree of garbage aggregation on the calculation. When the garbage shows an aggregated distribution, its cleaning difficulty is different from that in a dispersed distribution. The system analyzes the position information of the garbage in the image to judge the aggregation situation of the garbage. If it is found that the distance between the garbage in a certain area is less than the set threshold, it is considered that these garbage are in an aggregated state. For the aggregated garbage, an aggregation coefficient b is introduced when calculating the garbage equivalent. This coefficient is greater than 1, so that the equivalent value of the aggregated garbage increases relatively to reflect the increase in its cleaning difficulty. At this time, the calculation formula of the road surface garbage equivalent information is adjusted to: Through the above method, the operation optimization system can comprehensively consider various factors and accurately calculate the road surface garbage equivalent information.

[0040] In some embodiments, the washing and sweeping vehicle is equipped with various types of sensors to obtain comprehensive and accurate weather condition information and road condition information. The weather sensors include temperature and humidity sensors, rainfall sensors, wind speed and direction sensors, etc. The temperature and humidity sensor senses the temperature and humidity of the air through the sensing element. The rainfall sensor measures the rainfall using the electrical signal generated by the impact of raindrops. The wind speed and direction sensor determines the wind speed and direction through the rotation speed and direction of the wind cup or propeller. These sensors are installed at a high place on the washing and sweeping vehicle, such as the roof, to avoid being blocked and ensure that the weather condition can be monitored in real time and accurately. The road condition sensors include road surface humidity sensors, flatness sensors and image recognition sensors. The road surface humidity sensor measures the road surface humidity by sensing the moisture content of the road surface using the capacitance or resistance principle. The flatness sensor detects the undulation of the road surface through technologies such as laser ranging to judge the flatness of the road surface. The image recognition sensor is based on the road surface images collected by the camera and uses image recognition algorithms to identify whether there are potholes, cracks and other damages on the road surface. The road surface humidity sensor is installed at a position close to the road surface of the washing and sweeping vehicle to accurately measure the road surface humidity; the flatness sensor and the image recognition sensor are installed at the bottom of the washing and sweeping vehicle and can monitor the road surface that has been traveled in real time.

[0041] A garbage equivalent correction coefficient mapping database is preset, which stores the garbage equivalent correction coefficients corresponding to different combinations of weather and road conditions. The construction of the database is based on a large amount of experimental data and actual operation experience. In the experimental stage, various weather conditions are simulated, such as sunny days, rainy days, snowy days, etc., and different road conditions, such as dry roads, wet roads, damaged roads, etc., are set, and different garbage types and distributions are also set. Under each simulated scenario, indicators such as garbage weight and cleaning difficulty are measured to determine the corresponding correction coefficients. For example, in the case of rainy days and high road surface humidity, the garbage is easily soaked by rainwater and its weight increases, and at the same time it will stick to the road surface, resulting in increased cleaning difficulty. By measuring the increase ratio of garbage weight and the extension of cleaning time under different humidity thresholds in the experiment, combined with the increase degree of resource consumption in actual operation, the corresponding correction coefficient is determined. The data in the database is classified and stored according to weather conditions (such as weather type, rainfall, humidity range, etc.) and road conditions (such as road surface humidity, flatness, damage condition, etc.) to form a clear mapping relationship, which is convenient for the system to quickly query and call.

[0042] When the operation optimization system receives real-time weather condition information and road condition information from multiple sensors, it matches this information with the garbage equivalent correction coefficient mapping database. If the sensors detect that it is currently a rainy day and the road surface humidity measured by the road surface humidity sensor is greater than the preset humidity threshold, the system determines that there is a situation where rainwater increases the garbage weight and the garbage sticks to the road surface. At this time, based on the current weather and road condition information, the system searches for the corresponding correction coefficient in the garbage equivalent correction coefficient mapping database. Suppose the current road surface humidity on a rainy day is 80% and the preset humidity threshold is 60%. The system finds that the correction coefficient corresponding to this humidity range in the database is 1.3. The system applies the obtained correction coefficient to the road surface garbage equivalent information, multiplying the original road surface garbage equivalent information by the correction coefficient 1.3, thereby increasing the road surface garbage equivalent information. In this way, during subsequent operations, the road sweeper can reasonably adjust the working parameters of the roller brush assembly motor and the high-pressure water pump according to the corrected garbage equivalent information, increase the cleaning intensity, and cope with the increased garbage cleaning difficulty caused by weather and road conditions to ensure the cleaning effect.

[0043] S105. Adjust the working parameters of the roller brush assembly motor and the high-pressure water pump according to the set adjustment strategy based on the road surface garbage equivalent.

[0044] The operation optimization system will create a mapping relationship between the road surface garbage equivalent and the working parameters of the rotary brush assembly motor and the high-pressure water pump in advance through a large number of experiments and field tests. In the experimental stage, different road surface garbage conditions are simulated. For example, various types, different volumes, and different distribution densities of garbage are placed on the test road surface to create different road surface garbage equivalent environments. For each simulated situation, the working parameters of the rotary brush assembly motor and the high-pressure water pump are carefully adjusted, and their cleaning effects are observed to find the combination of working parameters of the rotary brush assembly motor and the high-pressure water pump that can achieve the best cleaning effect under different road surface garbage equivalents. These combinations include parameters such as the rotation speed of the rotary brush, the pressure of the rotary brush, the pressure of the water pump, and the water flow rate. These data are sorted out and stored in the system to form a detailed mapping table. When the system calculates the road surface garbage equivalent during actual operation, it will find the corresponding combination of working parameters for the current road surface garbage equivalent according to this mapping table. For example, if the road surface garbage equivalent calculated by the system is within a certain specific range, it will look up the corresponding rotation speed, pressure of the rotary brush assembly motor, and the pressure and water flow rate parameters of the high-pressure water pump from the mapping table.

[0045] The system collects the operation information of the rotary brush assembly and the high-pressure water pump in real time through sensors installed on the rotary brush assembly and the high-pressure water pump, such as rotation speed sensors, pressure sensors, etc. These sensors can monitor the actual rotation speed, pressure of the rotary brush, and the actual pressure and water flow rate of the water pump. Once these real-time data are received, the system will compare them with the set working parameters. If it is found that the rotation speed of the rotary brush does not reach the predetermined value, the system will adjust its operation through the controller of the rotary brush assembly motor. For example, if the rotation speed of the rotary brush is too low, the system will send a signal to the controller of the motor to increase the power supply to the motor, which may be to increase the current or voltage, so as to increase the rotation speed of the rotary brush to the predetermined level. For the high-pressure water pump, it is adjusted according to its type. If it is a variable-frequency water pump, the system will change the operation state of the water pump by adjusting the frequency of the frequency converter according to the feedback of the real-time pressure and water flow rate. When it is found that the pressure of the water pump is too high, the system will reduce the frequency of the frequency converter, thereby reducing the rotation speed of the water pump, so that the water flow rate and pressure can reach the expected level; conversely, if the pressure is too low, the frequency of the frequency converter will be increased to improve the performance of the water pump.

[0046] In the embodiments of the present application, by using the technical means of obtaining road surface image information based on multiple garbage monitors, determining garbage category information using a garbage classification model trained by deep learning, combining the image information to obtain geometric dimensions, garbage distribution density, and garbage volume information, determining weight coefficient information through a garbage category weight model, and then comprehensively determining the road surface garbage equivalent information, and adjusting the working parameters of the roller brush assembly motor and the high-pressure water pump according to the set adjustment strategy based on this information, it is possible to accurately analyze the road surface garbage condition and correspondingly adjust the operation parameters of the road sweeper, effectively solving the problems of inaccurate adjustment of cleaning parameters and difficulty in adapting to complex and changeable garbage conditions in the existing operation of road sweepers. Furthermore, the technical effects of strong pertinence, high cleaning efficiency, less resource waste, and good operation effect of the road sweeper cleaning operation are achieved.

[0047] In some embodiments, before the road sweeper operates, the operation optimization system calculates the expected garbage cleaning effect according to the set adjustment strategy, which is formulated based on a large number of experiments and historical operation data and covers the working parameter combinations of the roller brush assembly motor and the high-pressure water pump under different road surface garbage equivalents. For example, when the road surface garbage equivalent is in a certain range, the roller brush rotation speed is set to n revolutions per minute, and the high-pressure water pump pressure is set to p pascals. According to these parameters and information such as the cleaning speed and cleaning width of the road sweeper, combined with the distribution density and volume information of the garbage, a mathematical model is used to calculate the garbage equivalent that the sweeper is expected to clean up within a set time (such as 30 minutes). After the road sweeper completes the cleaning operation within the set time, the system obtains the actual garbage cleaning effect through various methods. A garbage collection amount detection device is installed on the road sweeper. For example, a weighing sensor is installed at the bottom of the garbage collection box, which can measure the weight change of the garbage in the collection box in real time. At the same time, using image recognition technology, the road surface is photographed before and after cleaning. By comparing the remaining garbage situation in the images, the uncleaned garbage is identified and its equivalent is estimated. For example, through image analysis software, information such as the category and area of the remaining garbage on the road surface after cleaning is identified, combined with the previously established relationship model between garbage volume and area, the volume of the remaining garbage is estimated, and then according to the garbage density information, the weight of the remaining garbage is calculated, so as to obtain the actual garbage equivalent cleaned up, that is, the actual garbage cleaning effect.

[0048] The system pre-constructs a weight adjustment model based on deep learning technology. During the model training phase, the system collects a large number of pairs of actual garbage cleaning effects and expected garbage cleaning effect data that annotate weight adjustment information. These data cover different operation scenarios, including different road surface conditions, garbage types and quantities, weather conditions, etc. The data is divided into a training set, a validation set, and a test set. The training set is used to train the model, the validation set is used to adjust the model hyperparameters to prevent overfitting, and the test set is used to evaluate the generalization performance of the model. During the training process, the training set is input into the model, and the model extracts and analyzes the features of the input data through a multi-layer neural network, and outputs a prediction of the weight adjustment information. When the actual garbage cleaning effect is less than the expected garbage cleaning effect, the system inputs these two data into the trained weight adjustment model. The model analyzes and calculates based on the input data and outputs weight adjustment information. This weight adjustment information contains weight adjustment suggestions for different garbage categories. For example, for a certain garbage category, the model may suggest increasing or decreasing its weight in the garbage category weight model by a certain proportion to reflect the difference between the difficulty of this type of garbage in the actual cleaning process and the expectation. According to the weight adjustment information output by the weight adjustment model, the operation optimization system adjusts the garbage category weight model. The garbage category weight model determines the relative importance of different garbage categories when calculating the road surface garbage equivalent. For example, if the weight adjustment information suggests increasing the weight coefficient of plastic garbage by 0.2, the system will find the corresponding weight coefficient of plastic garbage in the garbage category weight model and increase it by 0.2. In this way, the garbage category weight model can be dynamically adjusted according to the difference between the actual cleaning effect and the expectation, so that when calculating the road surface garbage equivalent subsequently, it can more accurately reflect the actual impact of various types of garbage on the sweeping operation, and further provide a more reasonable basis for adjusting the working parameters of the washing and sweeping vehicle, continuously optimizing the operation effect of the washing and sweeping vehicle, and improving the cleaning efficiency and resource utilization efficiency.

[0049] In some embodiments, the operation optimization system can collect a large amount of garbage category information obtained during the operation of the road sweeper and washer over a period of time. This information includes the time when the garbage is identified, the geographical location, and the category to which the garbage belongs. For the time dimension, the system divides a day into multiple time periods, such as in units of hours or half-hours. For the space dimension, the urban area is divided according to streets, communities, etc. to form multiple geographical area units. The system uses a spatio-temporal clustering algorithm to analyze the garbage category information. During the analysis process, the algorithm clusters the information that is close in time and space and has similar garbage categories into one category. For example, in a commercial area, during the period from 12:00 to 14:00 every day, the frequency of appearance of fast food boxes, beverage bottles and other garbage is relatively high, and the algorithm will cluster this information into one category, indicating that the distribution of such garbage is concentrated in this area during this period. In this way, the system obtains the garbage category distribution characteristics in different regions and different time periods. For example, domestic garbage appears more in the residential area in the morning, and paper garbage is more in the office area during the day on weekdays. Based on the obtained garbage category distribution characteristics, the operation optimization system uses machine learning algorithms, such as long short-term memory network (LSTM), to establish a garbage generation prediction model. When training the model, the historical garbage category distribution characteristic data, including the garbage categories and quantities in different regions at different time periods, are used as input data, and the corresponding time and region information are marked at the same time. The model learns the time series information and spatial characteristics in the input data through the LSTM network to mine the rules of garbage generation. After being trained with a large amount of data, the model can predict the garbage categories and quantities generated in a given region and time period according to the input region and time period information. The operation optimization system continuously runs the garbage generation prediction model. When the model predicts that the garbage generation amount in a certain region in the future time period will exceed the preset threshold, the system will automatically trigger the cleaning strategy adjustment mechanism. For the adjustment of the cleaning frequency, if it is predicted that the garbage generation amount in a commercial area will increase significantly on weekends at night, the system will adjust the cleaning frequency in this area from the original once every 3 hours to once every 2 hours on weekends at night to ensure that the garbage can be cleaned in time and avoid garbage accumulation.

[0050] In terms of the cleaning intensity, if it is predicted that the garbage generation amount increases and includes more difficult-to-clean garbage, such as construction waste, etc., the system will instruct the road sweeper and washer to increase the rotation speed of the roller brush assembly motor and the pressure of the high-pressure water pump when operating in this area to improve the cleaning strength. In this way, the cleaning strategy can be actively adjusted according to the prediction results. This foresight enables the cleaning work to adapt to the garbage changes in advance. For example, before holidays or special events, the cleaning arrangements can be planned in advance to effectively cope with the increase in the garbage volume and ensure that the urban environment always remains clean and tidy.

[0051] After combining the above content, the following is a further and more specific process description of the method provided in this embodiment. Please refer to Figure 2 , which is another process schematic diagram of the road sweeper and washer operation optimization method based on intelligent recognition in the embodiments of the present application.

[0052] S201. Collect road surface temperature data and road surface material information; The operation optimization system controls the temperature sensor and material detection sensor mounted on the road sweeper to collect road surface temperature data and road surface material information.

[0053] In terms of temperature data collection, a high-precision thermistor temperature sensor is adopted. This sensor utilizes the characteristic that the resistance value of the thermistor changes with temperature to convert the temperature signal into an electrical signal. It features rapid response and can collect road surface temperature data at a frequency of multiple times per second during the operation of the road sweeper, ensuring the real-time nature of the data. Moreover, the accuracy of this sensor can reach ±0.5°C, enabling precise measurement of the road surface temperature. By placing the sensor in multiple standard environments with known temperatures, recording the electrical signals output by the sensor, and establishing a calibration curve based on the relationship between the standard temperature values and the electrical signals. During actual use, the operation optimization system corrects the collected electrical signals according to the calibration curve to obtain accurate road surface temperature data.

[0054] For the collection of road surface material information, the system is equipped with multiple types of material detection sensors. For example, a spectral analyzer is used to conduct spectral analysis on the road surface material. Road surfaces of different materials will generate reflected spectra with different characteristics when irradiated by light of specific wavelengths. The spectral analyzer emits light of multiple wavelengths and receives the spectral signals reflected by the road surface. Through the analysis of the reflected spectra, the operation optimization system can identify whether the road surface is made of asphalt, cement, brick, or other materials. At the same time, in order to further improve the accuracy of material identification, the system also combines image texture analysis technology. Using a high-resolution camera installed on the road sweeper to capture road surface images, and then analyzing the road surface texture through image recognition algorithms. For example, the texture of an asphalt road is relatively rough and has a certain granularity, the texture of a cement road is relatively flat and may have regular expansion joint textures, and the brick road has a unique brick splicing texture. By extracting and analyzing these texture features and corroborating with the spectral analysis results, the operation optimization system can more accurately determine the road surface material information.

[0055] S202. When it is detected that the road surface temperature data exceeds the preset temperature threshold, obtain road surface thermal imaging data; The operation optimization system continuously monitors the road surface temperature data transmitted by the temperature sensor. A reasonable preset temperature threshold is set inside the system, and the determination of this threshold is based on the statistical analysis of a large amount of experimental data of different road surface materials under different environmental conditions. For example, for an asphalt road surface, when the temperature exceeds 60°C, the adhesion characteristics of sticky garbage such as chewing gum will change, and the cleaning difficulty will also change accordingly. Therefore, the preset temperature threshold for the asphalt road surface is set at 60°C.

[0056] Once the operation optimization system detects that the road surface temperature data exceeds the corresponding preset temperature threshold, it immediately triggers the thermal imaging device mounted on the road sweeper to start working. This thermal imaging device adopts advanced infrared thermal imaging technology and converts the infrared radiation energy emitted by an object into a thermal image by detecting it. The thermal imaging device features high resolution and a wide viewing angle, and can quickly obtain large-area road surface thermal imaging data during the driving of the road sweeper. Its resolution can reach 0.1 °C, and it can clearly distinguish tiny temperature differences on the road surface, so as to accurately identify objects such as chewing gum residues that have different temperatures from the surrounding road surface. During the operation of the thermal imaging device, in order to ensure the image quality and data accuracy, the device is equipped with autofocus and image enhancement functions. The autofocus function uses a built-in distance sensor to continuously monitor the distance between the thermal imaging device and the road surface, and automatically adjusts the lens focal length according to the distance to ensure that the captured thermal imaging image is clear. The image enhancement function uses digital image processing algorithms to process the original thermal imaging data, enhance the contrast and clarity of the image, and highlight the temperature anomaly areas.

[0057] The road surface thermal imaging data obtained by the thermal imaging device is transmitted to the operation optimization system through a high-speed data transmission line. The operation optimization system stores and preliminarily analyzes these data in real time, and compares the thermal imaging image with the pre-established thermal imaging feature library of chewing gum residues. The feature library stores the thermal imaging features of chewing gum residues under different environmental conditions and with different adhesion degrees. Through comparison, the operation optimization system can quickly screen out the areas where chewing gum residues may exist, preparing for the next step of precise identification.

[0058] S203. Identify the chewing gum residues on the road surface based on the road surface thermal imaging data, and record the position information, area information, and adhesion degree information of the chewing gum residues in the set garbage database; After the operation optimization system receives the road surface thermal imaging data, it uses image processing and pattern recognition technologies to deeply analyze the data to identify the chewing gum residues on the road surface. First, an image segmentation algorithm is used to distinguish the chewing gum residues in the thermal imaging image from the background. Specifically, a threshold-based segmentation method can be used. According to the temperature difference between the chewing gum residues and the surrounding road surface in the thermal imaging image, a suitable temperature threshold is set. When the pixel temperature value in the thermal imaging image is higher or lower than this threshold, it is determined as a pixel of the chewing gum residue, thus realizing the preliminary segmentation of the chewing gum residue area. When determining the location information of the chewing gum residue, the operation optimization system calculates its geographical location on the actual road surface through a coordinate conversion algorithm based on the pixel coordinates of the chewing gum residue area after segmentation, combined with the installation location, viewing angle of the thermal imaging device, and the driving trajectory information of the road sweeper. For example, if the thermal imaging device is installed at the center position of the bottom of the road sweeper, and the viewing angle of the device and the driving direction of the road sweeper are known, the pixel coordinates in the image can be converted into actual geographical coordinates through trigonometric relations, so as to accurately record the location information of the chewing gum residue.

[0059] Regarding the area information of the chewing gum residue, the operation optimization system estimates its area by calculating the number of pixels in the chewing gum residue area after segmentation. Since the resolution of the thermal imaging image is known, each pixel corresponds to a certain area on the actual road surface. By multiplying the number of pixels by the actual area corresponding to a single pixel, the approximate area of the chewing gum residue can be obtained.

[0060] In evaluating the adhesion degree of the chewing gum residue, the operation optimization system analyzes it in combination with the temperature change characteristics of the chewing gum residue and the heat conduction characteristics of the surrounding road surface. When the temperature of the chewing gum residue changes relatively slowly over a period of time, it indicates that its adhesion to the road surface is relatively tight, because the tightly adhered chewing gum residue can better conduct heat with the road surface, resulting in relatively stable temperature changes; conversely, if the temperature changes quickly, it indicates that the adhesion degree is relatively low. In addition, the operation optimization system can also assist in judging the adhesion degree by analyzing the edge blurring degree in the thermal imaging image of the chewing gum residue. A higher degree of edge blurring may mean a larger adhesion area between the chewing gum residue and the road surface and a higher adhesion degree. Finally, the operation optimization system records the location information, area information, and adhesion degree information of the identified chewing gum residues into the set garbage database. This database uses a relational database management system and can efficiently store and manage a large amount of special garbage information.

[0061] S204. Select the optimal cleaning agent formula from the cleaning agent formula database according to the road surface temperature data, the road surface material information, and the chewing gum residue information, and control the adjustment device to automatically adjust the ratio of the cleaning agent to water to obtain a special cleaning solution; When the operation optimization system executes this step, it first comprehensively analyzes the obtained road surface temperature data, road surface material information, and chewing gum residue information. The road surface temperature data provides the thermal conditions of the current road surface. At different temperatures, the effects of cleaning agents will vary. For example, at higher temperatures, the activity of some cleaning agents may increase, but they may also volatilize more quickly; the road surface material information determines the applicability of the cleaning agent because different materials have different adsorption and reaction characteristics with the cleaning agent. For example, asphalt roads are relatively porous and may be more likely to adsorb the cleaning agent, while cement roads are relatively hard and require a stronger cleaning ability; the chewing gum residue information includes its location, area, adhesion degree, etc. These information will affect the strength and action time of the cleaning agent required for cleaning.

[0062] To select the optimal cleaning agent formula from the cleaning agent formula database, which stores a large amount of cleaning agent formula information, and each formula is associated with different applicable conditions. This information is matched with the temperature range of the road surface, material type, and the state of chewing gum residue through associated keyword fields. For example, for chewing gum residues on an asphalt road with a medium adhesion degree and a temperature between 20°C and 30°C, the system will search the cleaning agent formula records in the database that meet these conditions. The system will sort the formulas according to the matching degree and select the formula with the highest matching degree as the optimal formula.

[0063] In terms of the operation optimization system automatically adjusting the ratio of the cleaning agent to water by the control and adjustment device, it is achieved through the connected proportional adjustment valve and metering pump. The proportional adjustment valve is a high-precision electric valve that can accurately control the flow rates of the cleaning agent and water. The metering pump can accurately measure the flow rates of the cleaning agent and water. Its principle is to accurately control the rotation speed and working time of the pump to deliver a certain volume of the cleaning agent and water to the mixing chamber. The system sends control signals to the proportional adjustment valve and metering pump according to the optimal ratio required by the selected cleaning agent formula. Assuming that the optimal cleaning agent formula requires a ratio of cleaning agent to water of 1:10, the system will calculate the volumes of the cleaning agent and water required for a specific cleaning task, and then send the corresponding flow signals to the proportional adjustment valve and metering pump to make them deliver at the corresponding flow rates and times.

[0064] To ensure the accuracy of the ratio, a concentration sensor is installed in the mixing chamber. This sensor determines whether the actual ratio meets the requirements by measuring certain physicochemical properties (such as conductivity, refractive index, etc.) of the mixed liquid. If the ratio deviates, the system will automatically adjust the working parameters of the proportional adjustment valve and metering pump until the preset ratio requirement is achieved. At the same time, to ensure the uniformity of mixing, a stirring device is set in the mixing chamber. This device can be a mechanical stirring paddle or an ultrasonic stirrer, which makes the cleaning agent and water fully mixed through high-speed stirring or ultrasonic vibration to form a special cleaning liquid.

[0065] S205. Monitor the road surface temperature change in real time through a thermal sensor. When the temperature reaches the set temperature range suitable for removing chewing gum residues, control the road sweeper to use a special cleaning liquid for a special cleaning procedure.

[0066] The operation optimization system monitors the road surface temperature change in real time through a thermal sensor. The thermal sensors are distributed at different positions of the road sweeper, such as around the cleaning brush heads and near the spraying devices, etc., to ensure comprehensive monitoring of the road surface temperature. These thermal sensors use highly sensitive thermistors or thermocouples, which can quickly sense the subtle changes in the road surface temperature, convert the temperature signal into an electrical signal, and transmit it to the operation optimization system.

[0067] The temperature range suitable for removing chewing gum residues is preset in the system, and this range is determined through a large number of experiments and actual cleaning experiences. For example, for a certain combination of a specific cleaning agent and chewing gum residues, the cleaning effect may be best within the temperature range of 30°C to 40°C. When the temperature signal transmitted by the thermal sensor shows that the road surface temperature enters this set temperature range, the operation optimization system will trigger the corresponding control logic.

[0068] When controlling the road sweeper to use a special cleaning liquid for a special cleaning procedure, the system will first open the valve of the storage tank of the special cleaning liquid, start the delivery pump of the special cleaning liquid, and transport the cleaning liquid to the nozzle through a pipeline. The design and layout of the nozzle are carefully designed according to the cleaning range and cleaning effect requirements of the road sweeper. It can be an array composed of multiple nozzles, covering a certain width and angle, ensuring that the cleaning liquid can be evenly sprayed on the chewing gum residues. The opening and closing of the nozzle are controlled by an electromagnetic valve, and the system can accurately control the working state of the nozzle according to needs, such as the opening time and spraying intensity, etc.

[0069] During the special cleaning procedure, it will also control other cleaning components of the road sweeper to cooperate. For example, adjust the rotation speed and pressure of the rotary brush to match the cleaning effect of the cleaning liquid. For different adhesion degrees of chewing gum residues, different settings can be made for the rotation speed and pressure of the rotary brush. When the adhesion degree is high, the rotary brush works at a higher rotation speed and pressure to enhance the cleaning effect; while for the case of low adhesion degree, a lower rotation speed and pressure are adopted to avoid unnecessary wear on the road surface.

[0070] In the embodiments of the present application, since a series of closely connected and highly targeted technical means are adopted, from collecting road surface temperature and material information, to obtaining thermal imaging data based on the temperature to identify chewing gum residues, and then to formulating a special cleaning solution based on this and cleaning at an appropriate temperature, it is possible to accurately locate chewing gum residues and adopt appropriate methods for treatment, effectively solving the problems of difficult effective removal of special garbage such as chewing gum and poor cleaning targeting in the operation of existing washing and sweeping vehicles. Furthermore, it realizes efficient cleaning of special garbage, avoids the influence of chewing gum residues on the road surface beauty and cleanliness, and improves the technical effect of the intelligent and precise level of the washing and sweeping vehicle in dealing with special garbage scenarios.

[0071] In some embodiments, an odor sensor array can be installed on the washing and sweeping vehicle. The array is composed of various different types of odor sensors, and each sensor has high sensitivity to specific odor molecules. These sensors are reasonably distributed at the front end, both sides, etc. of the washing and sweeping vehicle, and can collect road surface odor data in all directions. When the washing and sweeping vehicle is driving, the odor sensor array real-time senses odor molecules in the surrounding air and converts them into electrical signals and transmits them to the operation optimization system.

[0072] The operation optimization system pre-stores the odor characteristic data emitted by special garbage such as pet feces and rotting organic matter. The system analyzes and processes the received odor electrical signals and compares them with the preset odor characteristics. For example, pet feces contain specific volatile organic compounds to which the odor sensors are sensitive. By identifying the corresponding electrical signal patterns, the system determines whether there is pet feces or rotting organic matter in the target area. If the matching degree between the collected odor data and the preset odor characteristics of pet feces exceeds the set threshold, the system determines that there is such special garbage in the area. When the odor sensor array detects the set odor, the operation optimization system immediately activates the ultraviolet spectral analyzer installed on the washing and sweeping vehicle. The ultraviolet spectral analyzer emits ultraviolet rays to irradiate the special garbage in the target area. After the substances in the special garbage absorb the ultraviolet rays, they will produce specific spectral characteristics. The analyzer collects and analyzes these spectral data to determine the material composition of the garbage. Based on the results of the material composition analysis, the operation optimization system selects a dedicated bio-enzyme degrading agent from the pre-established bio-enzyme degrading agent database. The database stores information on bio-enzyme degrading agents for different material compositions, and each degrading agent has its applicable garbage types and composition ranges. For example, for pet feces containing more protein, the system will select a bio-enzyme degrading agent that can efficiently decompose protein; for rotting plant organic matter rich in carbohydrates, a corresponding carbohydrate degrading enzyme preparation is selected. The operation optimization system controls the micro-injection system on the washing and sweeping vehicle to release the selected bio-enzyme degrading agent to the target area. The micro-injection system evenly sprays the bio-enzyme degrading agent on the special garbage through high-precision nozzles. To monitor the degree of chemical reaction during the degradation process in real time, the system uses a variety of monitoring means, such as installing chemical sensors near the spraying area to detect the change in the concentration of specific chemical substances generated during the degradation process; or using spectral analysis technology to monitor the change in the material composition of the garbage in real time.

[0073] According to the monitored degree of chemical reaction, the operation optimization system dynamically adjusts the spraying amount of the enzyme preparation. If the chemical reaction rate is slow, indicating poor degradation effect, the system will increase the spraying frequency or dosage of the micro-injection system to accelerate the degradation process; conversely, if the chemical reaction is too intense, it may cause waste of resources or other adverse effects, and the system will appropriately reduce the spraying amount to ensure the efficiency and stability of the degradation process. After the bio-enzyme degradation treatment is completed, to eliminate the residual odor, the operation optimization system activates the ozone generator. The ozone generator generates ozone gas, which is transported through pipelines to the target area for local air purification. Ozone has strong oxidizing properties and can chemically react with the residual odor molecules to decompose them into harmless substances, thereby effectively eliminating the odor. This can timely handle special garbage and eliminate odors, maintaining the cleanliness and hygiene of urban roads.

[0074] The operation optimization system in the embodiment of the present invention application will be described from the perspective of hardware processing below. Please refer toFigure 3 , which is a schematic structural diagram of an entity device of the operation optimization system in the embodiment of the present application.

[0075] It should be noted that Figure 3 the structure of the operation optimization system shown is only an example, and should not bring any limitations to the functions and usage scope of the embodiments of the present invention.

[0076] As Figure 3 shown, the operation optimization system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage section 308 into the random access memory (RAM) 303, such as executing the method described in the above embodiments. In the RAM 303, various programs and data required for system operation are also stored. The CPU 301, ROM 302, and RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0077] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD), an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed, so that a computer program read from it can be installed into the storage section 308 as needed.

[0078] Specifically, according to the embodiments of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments of the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication section 309, and / or installed from the removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are executed.

[0079] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0080] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. Among them, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the above-mentioned module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the block may occur in a different order than that marked in the accompanying drawings.

[0081] Specifically, the job optimization system of this embodiment includes a processor and a memory. A computer program is stored on the memory. When the computer program is executed by the processor, it implements the method for optimizing the operation of a washing and sweeping vehicle based on intelligent recognition provided in the above-mentioned embodiment.

[0082] On the other hand, the present invention also provides a computer-readable storage medium. This storage medium can be included in the job optimization system described in the above-mentioned embodiment; or it can exist independently and not be assembled into the job optimization system. The above-mentioned storage medium carries one or more computer programs. When the one or more computer programs are executed by a processor of the job optimization system, the job optimization system is enabled to implement the method for optimizing the operation of a washing and sweeping vehicle based on intelligent recognition provided in the above-mentioned embodiment.

[0083] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the various embodiments of the present application.

[0084] As used in the foregoing embodiments, depending on the context, the term "when" may be construed to mean "if", "after", "in response to determining", or "in response to detecting". Similarly, depending on the context, the phrase "when determining" or "if detecting (the stated condition or event)" may be construed to mean "if determining", "in response to determining", "when detecting (the stated condition or event)", or "in response to detecting (the stated condition or event)".

[0085] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the foregoing embodiments can be implemented by a computer program instructing relevant hardware. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the foregoing method embodiments. The foregoing storage media include various media that can store program codes, such as ROM, random access memory (RAM), magnetic disks, or optical discs.

Claims

1. A method for optimizing the operation of a washing and sweeping vehicle based on intelligent identification, applied to an operation optimization system, characterized in that: The method comprises: obtaining road surface image information through a plurality of garbage monitors; In combination with the road surface image information, garbage classification information is determined through a garbage classification model, wherein the garbage classification model is obtained by deep learning training of garbage image data with multiple labeled category information in advance; Determine the geometrical dimension data and the garbage distribution density information corresponding to the garbage category information in combination with the road surface image information, and determine the garbage volume information through the geometrical dimension data; In combination with the garbage category information, a plurality of weight coefficient information is determined according to a garbage category weight model, wherein the garbage category weight model is obtained in advance by deep learning training based on a plurality of garbage image data with labeled weight coefficients; Determining road garbage equivalent information by combining the garbage volume information, the weight coefficient information and the distribution density information; The operating parameters of the roller brush assembly motor and the high-pressure water pump are adjusted according to the set adjustment strategy based on the road garbage equivalent.

2. The method according to claim 1, characterized in that: After the step of determining the road garbage equivalent information by combining the garbage volume information, the weight coefficient information and the distribution density information, the step further includes: Determine weather condition information and road condition information through multiple sensors; In combination with the weather condition information and the road condition information, the road surface garbage equivalent information is corrected according to a preset garbage equivalent correction coefficient mapping database, and the correction at least includes when the sensor detects that it is currently raining and the road surface humidity is greater than a preset humidity threshold, it is determined that there is rain water that increases the weight of the garbage and the garbage is stuck to the road surface, then the corresponding correction coefficient is obtained from the garbage equivalent correction coefficient mapping database to increase the road surface garbage equivalent information.

3. The method according to claim 1, characterized in that After the step of determining the road garbage equivalent information by combining the garbage volume information, the weight coefficient information and the distribution density information, the step further includes: If the road garbage equivalent information is greater than a preset equivalent threshold, a warning message is sent to the user terminal, and the cleaning equipment is started according to the preset additional cleaning plan to enhance the cleaning measures.

4. The method according to claim 1, characterized in that: After the step of adjusting the working parameters of the roller brush assembly motor and the high-pressure water pump according to the set adjustment strategy based on the road surface garbage equivalent, the method further includes: Obtaining an expected garbage cleaning effect within a set time, wherein the expected garbage cleaning effect refers to the amount of garbage equivalent that is expected to be cleaned by the cleaning vehicle within the set time after adjustment according to the set adjustment strategy; Obtaining the actual garbage cleaning effect within the set time; If the actual garbage cleaning effect is less than the expected garbage cleaning effect, the actual garbage cleaning effect and the expected garbage cleaning effect are used to determine weight adjustment information through a weight adjustment model, wherein the weight adjustment model is obtained by deep learning training in advance based on actual garbage cleaning effects and expected garbage cleaning effects with multiple labeled weight adjustment information; The garbage category weight model is adjusted according to the weight adjustment information.

5. The method according to claim 1, characterized in that After the step of determining the garbage category information through the garbage classification model, the step further includes: Performing spatiotemporal clustering analysis on the garbage category information to obtain garbage category distribution characteristics in different regions and time periods; Establishing a garbage generation prediction model based on the garbage category distribution characteristics, wherein the prediction model is used to predict the type and amount of garbage generated in a set area during a set period of time; When the prediction results show that the amount of garbage generated in a certain area in the future period will exceed the preset threshold, the cleaning frequency and intensity of the area will be adjusted in advance.

6. The method according to claim 1, characterized in that After the step of determining the garbage category information by using the garbage classification model in combination with the road surface image information, the method further includes: Collect road surface temperature data and road surface material information; When it is detected that the road surface temperature data exceeds a preset temperature threshold, obtaining road surface thermal imaging data; Identifying chewing gum residues on a road surface based on the road surface thermal imaging data, and recording the location information, area information, and adhesion degree information of the chewing gum residues into a set garbage database; According to the road surface temperature data, the road surface material information and the chewing gum residue information, an optimal detergent formula is selected from a detergent formula database, and a regulating device is controlled to automatically adjust the ratio of the detergent to water to obtain a special cleaning solution; The temperature change of the road surface is monitored in real time through thermistor sensors. When the temperature reaches the set temperature range suitable for removing chewing gum residues, the washing and sweeping vehicle is controlled to use special cleaning fluid to perform a special cleaning procedure.

7. The method according to claim 1, characterized in that After the step of determining a plurality of weight coefficient information according to the garbage category weight model, the method further includes: Collect road odor data through an odor sensor array; Determining whether there is pet feces or decayed organic matter in the target area based on the road odor data; When the set odor is detected, the UV spectrometer is started to analyze the material composition, and a special bio-enzyme degradation agent is selected according to the material composition analysis results; Controlling the micro-injection system to release the bio-enzyme degradation agent to the target area; Real-time monitoring of the degree of chemical reaction during the degradation process and dynamic adjustment of the spraying amount of enzyme preparation; After the degradation process is completed, local air purification is carried out through an ozone generator to eliminate residual odors.

8. A job optimization system, characterized in that: The job optimization system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the job optimization system to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on a job optimization system, the job optimization system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is executed on a job optimization system, the job optimization system is caused to execute the method according to any one of claims 1 to 7.

Images