An image recognition-based sweeper brush control method and system, and a medium

Abstract

The present application relates to the technical field of image processing, in particular to a cleaning vehicle cleaning brush control method and system based on image recognition and a medium, the method comprising: acquiring a road surface image, and extracting a bounding box of each type of garbage in the road image; determining a garbage type probability distribution of pixels in each bounding box, and calculating a cross degree of each type of garbage by combining the garbage type probability distribution and the pixel position of each type of garbage; establishing a pollution weight of the bounding box based on the garbage type of each pixel, determining a garbage concentration of each bounding box based on the cross degree and the pollution weight of each type of garbage in each bounding box; performing clustering analysis on each bounding box based on the garbage concentration and the position distribution of each bounding box, and obtaining a cleaning area; determining a cleaning difficulty of the cleaning area based on the garbage concentration of each bounding box in the cleaning area, and controlling the angle and rotating speed of the cleaning brush in each cleaning area according to the cleaning difficulty; the present application can realize efficient and accurate cleaning, and improve the cleaning efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a cleaning brush control method and system for a cleaning vehicle based on image recognition and a medium. BACKGROUND

[0002] When cleaning the road surface, the cleaning vehicle often cannot accurately identify the distribution of garbage, resulting in low cleaning efficiency. Traditional methods rely on human experience and lack intelligent control, making it difficult to adapt to complex cleaning environments.

[0003] In related technologies, image recognition technology has been preliminarily applied in garbage identification and cleaning path planning, but intelligent adjustment of the angle and speed of the cleaning brush has not been achieved, resulting in unstable cleaning effect. SUMMARY

[0004] To solve the above problems, the present application provides a cleaning brush control method and system for a cleaning vehicle based on image recognition and a medium, which aims to automatically analyze the distribution and concentration of garbage through image recognition technology, intelligently adjust the angle and speed of the brush, achieve efficient and accurate cleaning, and improve cleaning efficiency.

[0005] To achieve the above purpose, the present application provides the following technical solutions:

[0006] On the one hand, the present application provides a cleaning brush control method for a cleaning vehicle based on image recognition, which comprises the following steps:

[0007] Obtain the road surface image, and use a target detection model to extract the bounding box of each type of garbage in the road image;

[0008] Determine the garbage type probability distribution of each pixel in the bounding box, and calculate the intersection degree of each type of garbage in the bounding box based on the garbage type probability distribution and the pixel position of each type of garbage;

[0009] Establish the pollution weight of the bounding box based on the garbage type of each pixel in the bounding box, and determine the garbage concentration of each bounding box based on the intersection degree and pollution weight of each type of garbage in each bounding box;

[0010] Cluster analysis is performed on each bounding box based on the garbage concentration and position distribution of each bounding box, and the cleaning area is obtained;

[0011] Determine the cleaning difficulty of the cleaning area based on the garbage concentration of each bounding box in the cleaning area, and control the angle and speed of the cleaning brush in each cleaning area according to the cleaning difficulty.

[0012] Optionally, the determination of the garbage type probability distribution of each pixel in the bounding box, and the calculation of the intersection degree of garbage in the bounding box based on the garbage type probability distribution and the pixel position of each type of garbage, comprises:

[0013] The frequency of each pixel in the bounding box belonging to the target type garbage is accumulated, and the total number of pixels is obtained. The probability of the target type garbage in the bounding box is obtained.

[0014] The pixel position difference of the target type garbage in the two bounding boxes is calculated, and the pixel difference of the target type garbage in the two bounding boxes is obtained.

[0015] The probability and pixel difference of each target type garbage in the two bounding boxes are multiplied and accumulated, and the intersection degree of the target type garbage in the two bounding boxes is obtained. The intersection degree of each type garbage in the bounding box is accumulated.

[0016] Optionally, the pollution weight of the bounding box is established based on the garbage type of each pixel in the bounding box, and the garbage concentration of each bounding box is determined based on the intersection degree and pollution weight of each type garbage in each bounding box, comprising:

[0017] According to the basic weight coefficient of the garbage type, the garbage concentration of each type garbage in the bounding box is multiplied by the corresponding basic weight coefficient, and the pollution weight of each type garbage is obtained.

[0018] The pollution weight of each type garbage in the bounding box is weighted and summed with the intersection degree, and the comprehensive pollution index of the bounding box is obtained.

[0019] Based on the comprehensive pollution index of the bounding box and the area of the bounding box, the garbage concentration value per unit area is calculated as the garbage concentration of the bounding box.

[0020] Optionally, the garbage concentration and position distribution of each bounding box are analyzed to obtain a cleaning area, comprising:

[0021] The cleaning radius is set according to the coverage area of the cleaning brush and the average value of the garbage concentration of each bounding box, the minimum point number is set according to the minimum garbage number in the road surface image as a coverage area, and the clustering center is determined based on the garbage concentration of each bounding box.

[0022] The cleaning radius is used as the clustering boundary of the clustering center, and the adjacent bounding boxes of the clustering center are clustered within the minimum point number to form a cleaning area.

[0023] The cleaning area is equally divided into multiple sub-areas, and the local garbage concentration of the sub-area is determined based on the area proportion and garbage concentration of each bounding box in the sub-area.

[0024] The distance of each sub-area and the clustering center is determined, the sub-area with the maximum product of local garbage concentration and distance is selected as the new clustering center, and it is determined whether the distance change of the clustering center is lower than the distance threshold. If yes, S450 is executed; if no, S420 is executed.

[0025] Gradient calculation is performed on the garbage concentration of each bounding box in the cleaning area, a plurality of bounding boxes with large gradients are taken as reference boxes, an external curve containing the plurality of reference boxes is generated, and a cleaning area containing the external curve is generated with the clustering center as the center and within the cleaning radius and the minimum point number.

[0026] Optionally, the clustering of the adjacent bounding boxes of the clustering center within the minimum point number to form the cleaning area includes:

[0027] The adjacent bounding boxes are screened according to the clustering center and the cleaning radius, and the number of the bounding boxes in each cluster is ensured to be not less than the minimum point number, thereby forming a cleaning unit.

[0028] If there is an overlapping bounding box on the edge of the cleaning unit, the bounding box is divided into a cleaning unit with a larger pixel ratio, the cleaning unit is adjusted in position based on the bounding box in the cleaning unit, and the adjusted cleaning unit is taken as the cleaning area.

[0029] Optionally, the garbage concentration of the cleaning area is determined, and the angle and the rotating speed of the cleaning brush in each cleaning area are controlled according to the garbage concentration.

[0030] The garbage concentration of each bounding box in the cleaning area is accumulated and then subjected to exponential normalization processing, thereby obtaining the cleaning difficulty of the cleaning area.

[0031] The range value of the bristle angle is obtained, a first functional relationship between the cleaning difficulty and the bristle angle is established within the range value, and the bristle angle is adjusted within the range value based on the cleaning difficulty.

[0032] The rotating speed of the cleaning brush, the driving speed of the cleaning vehicle, the reference speed of the cleaning vehicle and the basic rotating speed of the cleaning brush are obtained, and the rotating speed of the cleaning brush is determined in combination with the cleaning difficulty of the garbage.

[0033] On the other hand, the embodiment of the present application provides a cleaning brush control system of a cleaning vehicle based on image recognition, which comprises:

[0034] At least one processor;

[0035] At least one memory for storing at least one program;

[0036] When the at least one program is executed by the at least one processor, the at least one processor implements the above-mentioned method.

[0037] On the other hand, the embodiment of the present application provides a computer readable storage medium, wherein a processor executable program is stored, and the processor executable program is used to execute the above-mentioned method when executed by a processor.

[0038] The application discloses a cleaning vehicle cleaning brush control method and system based on image recognition and a medium. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0040] Figure 1 is a flow diagram of a cleaning vehicle cleaning brush control method based on image recognition according to an embodiment of the present application.

[0041] Figure 2 is a structural diagram of a cleaning vehicle cleaning brush control system based on image recognition according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] The disclosed concept, specific structure and generated technical effects will be described clearly and completely in the following embodiments and drawings, so as to fully understand the purpose, scheme and effect of the present application. It should be noted that the embodiments and features in the embodiments in the present application can be combined with each other without conflict.

[0043] The present application aims to realize intelligent control of the cleaning brush through image recognition technology, improve the cleaning efficiency, and reduce manual intervention. Through real-time image analysis, the angle and speed of the cleaning brush are automatically adjusted to ensure comprehensive and uniform cleaning coverage, improve the operation convenience, and realize efficient and intelligent cleaning.

[0044] REFERENCE Figure 1 For example, Figure 1A cleaning brush control method of a cleaning vehicle based on image recognition is shown, and the method comprises the following steps:

[0045] S100, acquiring a road surface image, and extracting a boundary box of each type of garbage in the road surface image by using a target detection model;

[0046] Specifically, the garbage features in the road surface image are extracted by using the trained target detection algorithm, the extracted features are classified to determine the type of garbage (such as plastic bags, paper scraps, cigarette butts, leaves, etc.) in the road surface image, and the position of the garbage is located. The boundary box of the garbage is output in the road surface image, and the target detection algorithm can use YOLO (You Only Look Once), Faster R-CNN, etc.

[0047] S200, determining the garbage type probability distribution of each pixel in the boundary box, and calculating the intersection degree of each type of garbage in the boundary box according to the garbage type probability distribution and the pixel position of each type of garbage;

[0048] Specifically, after obtaining the boundary box of all garbage in the road surface image, for the intersecting boundary box, it indicates that there is garbage accumulation, which needs to be handled specially and marked as a high-priority cleaning area. Subsequently, the rotation speed and the pressing degree of the cleaning brush are dynamically adjusted according to the garbage type and the accumulation degree to ensure complete removal. In this way, the cleaning vehicle can intelligently identify and efficiently handle various road surface garbage, improving the cleaning effect. For the boundary boxes with a long distance, they are regarded as sporadic garbage distribution areas, and the cleaning intensity is appropriately reduced to avoid resource waste.

[0049] S300, establishing a pollution weight of the boundary box based on the garbage type of each pixel in the boundary box, and determining the garbage concentration of each boundary box based on the intersection degree and the pollution weight of each type of garbage in each boundary box;

[0050] Specifically, a dynamically updated garbage concentration field is established based on the real-time collected road surface image, and the physical properties of different garbage types, the garbage concentration, and the mutual superposition relationship are comprehensively considered to provide accurate data support for the cleaning strategy.

[0051] The pollution weight is dynamically adjusted according to the garbage type, volume, and cleaning difficulty. For example, garbage such as cigarette butts and candy paper that is difficult to clean is given a higher weight, while fallen leaves and other garbage that is easy to clean is given a relatively lower weight value. By continuously optimizing the weight allocation strategy, the cleaning decision is more accurate and efficient.

[0052] In the subsequent cleaning process, for the high-concentration area, the rotation speed and the pressing degree of the cleaning brush are automatically increased; for the low-concentration area, the energy-saving mode is used to run, realizing the optimal allocation of resources.

[0053] S400, cluster analysis is performed on each bounding box based on the garbage concentration and position distribution of each bounding box, and a cleaning area is obtained;

[0054] Specifically, the spatial position data of each bounding box in the road surface image is input into the selected clustering algorithm. The clustering algorithm clusters the bounding boxes according to the set parameters, and divides the bounding boxes with similar spatial positions into the same cluster. Each cluster corresponds to a potential cleaning area.

[0055] S500, the cleaning difficulty of the cleaning area is determined based on the garbage concentration of each bounding box in the cleaning area, and the angle and rotating speed of the cleaning brush in each cleaning area are controlled according to the cleaning difficulty.

[0056] In the embodiments provided by the application, the road surface image is monitored in real time, and the cleaning strategy is dynamically adjusted to ensure that the garbage in each area can be effectively removed, thereby greatly improving the cleaning efficiency and road surface cleanliness.

[0057] As an improvement of the above-mentioned embodiments, in S200, the determination of the garbage type probability distribution of each pixel in the bounding box, the garbage type probability distribution and the pixel position of each type of garbage are combined to calculate the intersection degree of the garbage in the bounding box, comprising:

[0058] S210, the ratio of the frequency of each pixel belonging to the target type garbage to the total number of pixels in the bounding box is accumulated to obtain the probability of the target type garbage in the bounding box;

[0059] S220, the pixel position difference of the target type garbage in two bounding boxes is calculated to obtain the pixel difference of the target type garbage in the two bounding boxes;

[0060] S230, the probability and pixel difference of each target type garbage in the two bounding boxes are multiplied and accumulated to obtain the intersection degree of the target type garbage in the two bounding boxes, and the intersection degrees of each target type garbage in the bounding box are accumulated to obtain the intersection degrees of each type of garbage in the bounding box.

[0061] Specifically, the intersection degree of the garbage type in the bounding box is accurately calculated in combination with the position information, and the intersection degree calculation formula is:

[0062] Where P(i,j) is the intersection degree of the i th bounding box and the j th bounding box, p(k) is the probability of the k th type of garbage, and d(k,i,j) is the pixel difference of the k th type of garbage in the i th and j th bounding boxes.

[0063] d(k,i,j) is obtained by calculating the pixel position difference of each type of garbage in the two bounding boxes, which comprehensively evaluates the garbage distribution in the overlapping area and ensures the accuracy of the intersection degree calculation.

[0064] where f(m, k) is the frequency of the mth pixel belonging to the kth type of garbage, and T is the total number of pixels. Through this formula, the system can accurately evaluate the degree of intersection of garbage types within each bounding box, thereby optimizing the region division and cleaning strategy.

[0065] By classifying and identifying each pixel, the frequency of its occurrence in each type of garbage is counted, ensuring the accuracy of f(m, k). Combined with the total number of pixels T, the value of p(k) is accurately calculated, and then the degree of intersection is evaluated through the P(i, j) formula, the garbage region division is optimized, and the cleaning efficiency is improved. The ratio of f(m, k) to T reflects the distribution of the kth type of garbage in the entire image, directly affecting the calculation accuracy of the degree of intersection. Accurate f(m, k) and T values ensure the reliability of p(k), making the bounding box merging and region division more reasonable, and ultimately achieving efficient and accurate cleaning path planning.

[0066] As an improvement of the above embodiment, in S300, the pollution weight of the bounding box is established based on the garbage type of each pixel within the bounding box, and the garbage concentration of each bounding box is determined based on the degree of intersection and the pollution weight of each type of garbage within each bounding box, comprising:

[0067] S310, according to the garbage type, set the basic weight coefficient, multiply the garbage concentration of each type of garbage in the bounding box by the corresponding basic weight coefficient, to obtain the pollution weight of each type of garbage;

[0068] Specifically, according to the garbage type, set the basic weight coefficient, through real-time image analysis, calculate the garbage concentration of each type of garbage in the bounding box, multiply the garbage concentration by the basic weight coefficient, to obtain the dynamic pollution weight. The pollution weight can accurately reflect the influence degree of different types of garbage on the cleaning difficulty, and provide a reliable basis for subsequent garbage concentration calculation.

[0069] S320, weight sum the pollution weight of each type of garbage in the bounding box and its degree of intersection to obtain the comprehensive pollution index of the bounding box;

[0070] Specifically, by introducing the intersection degree correction coefficient, the pollution index of the overlapping area is nonlinearly amplified, ensuring that the pollution degree of the high garbage concentration accumulation area is fully reflected. At the same time, the linear weighting method is adopted for the isolated distribution of garbage area, keeping the rationality of the pollution index.

[0071] S330, based on the comprehensive pollution index of the bounding box and the area of the bounding box, calculate the garbage concentration value per unit area as the garbage concentration of the bounding box.

[0072] Specifically, by dividing the comprehensive pollution index by the area of the bounding box, area normalization processing is performed to eliminate the influence of the size of the bounding box on the concentration evaluation, and to ensure the comparability of the garbage concentration of regions of different sizes. The final generated garbage concentration field can accurately reflect the pollution degree of each region of the road surface, and provide a quantitative basis for intelligent cleaning decision-making.

[0073] As an improvement of the above embodiment, in S400, the garbage concentration and position distribution of each bounding box are subjected to clustering analysis to obtain a cleaning area, including:

[0074] In S410, a cleaning radius is set according to the coverage area of the cleaning brush and the average value of the garbage concentration of each bounding box, a minimum point number is set according to the minimum amount of garbage considered as one coverage area in the road surface image, and a clustering center is determined based on the garbage concentration of each bounding box;

[0075] Specifically, appropriate cleaning radius and minimum point number are set, and the setting of these two parameters needs to be adjusted according to the actual distribution of the bounding box and the expected coverage area division granularity. If the garbage distribution is relatively dense, the cleaning radius can be appropriately set small, and the minimum point number is set according to the minimum amount of garbage considered as one coverage area in the road surface image.

[0076] In S420, the cleaning radius is used as the clustering boundary of the clustering center, and the adjacent bounding boxes of the clustering center are clustered within the minimum point number to form a cleaning area;

[0077] Specifically, each bounding box is arranged in descending order of garbage concentration, and the first N high-concentration bounding boxes are selected as the clustering center. It can be understood that the clustering boundary divided in the area centered on the clustering center and with the cleaning radius as the radius contains multiple bounding boxes.

[0078] In S430, the cleaning area is evenly divided into multiple sub-areas, and the local garbage concentration of each sub-area is determined based on the area proportion and garbage concentration of each bounding box in the sub-area;

[0079] Specifically, the area proportion of each bounding box in the sub-area is determined, and the local garbage concentration of the sub-area is calculated by weighted average of each bounding box in the sub-area according to the area proportion and garbage concentration, to ensure the reasonable setting of the subsequent clustering center.

[0080] In S440, the distance of each sub-area and the clustering center is determined, the sub-area with the maximum product of local garbage concentration and distance is selected as the new clustering center, and it is determined whether the distance change of the clustering center is lower than the distance threshold value. If yes, S450 is executed; if no, S420 is executed;

[0081] Specifically, the product of the distance of the garbage concentration of each sub-region and the cluster center is calculated, the sub-region with the maximum product is selected as the new cluster center, and the cleaning efficiency is maximized. Through continuous iteration and optimization, the optimal cleaning area in the road image is finally determined, and efficient and intelligent cleaning is realized.

[0082] S450, the gradient of the garbage concentration of each bounding box in the cleaning area is calculated, and the multiple bounding boxes with larger gradients are taken as the reference boxes. An external curve containing multiple reference boxes is generated, and a cleaning area containing the external curve is generated with the cluster center as the center, within the cleaning radius and the minimum number of points.

[0083] By calculating the gradient change of the garbage concentration in the cleaning area, the larger the gradient change, the more obvious the boundary, and the more convenient it is to focus on cleaning. Based on the cluster center, the cleaning radius and the minimum number of points, the cleaning area boundary is accurately divided to avoid missing or repeated cleaning. Through iterative optimization of the clustering parameters, a comprehensive and clear boundary cleaning area division scheme is finally formed, laying a foundation for subsequent path planning.

[0084] As an improvement of the above embodiment, in S420, the cluster boundary with the cleaning radius as the cluster center clusters the adjacent bounding boxes of the cluster center within the minimum number of points to form a cleaning area, comprising:

[0085] S421, the adjacent bounding boxes are screened according to the cluster center and the cleaning radius to ensure that the number of bounding boxes in each cluster is not less than the minimum number of points to form a cleaning unit;

[0086] Specifically, by checking the bounding boxes around the cluster center one by one, the bounding boxes that do not meet the conditions such as being too far away or having too low garbage concentration are removed to ensure that the number of bounding boxes in each cleaning unit meets the minimum number of points requirement, thereby accurately dividing the cleaning area.

[0087] S422, if there are overlapping bounding boxes on the edge of the cleaning unit, the bounding boxes are divided into the cleaning unit with a larger pixel ratio, and the cleaning unit is adjusted in position based on the bounding boxes in the cleaning unit. The adjusted cleaning unit is taken as the cleaning area.

[0088] The formed cleaning unit is optimized and adjusted to ensure that the boundaries of each unit are clear and there is no overlap. By calculating the overlapping area of the bounding boxes of each cleaning unit, the boundary position is adjusted to ensure that each cleaning unit is independent and complete. Finally, an accurate cleaning area map is generated to guide the efficient operation of the intelligent cleaning vehicle.

[0089] As an improvement of the above embodiment, in S500, the garbage concentration of the cleaning area is determined, and the angle and speed of the cleaning brush in each cleaning area are controlled according to the garbage concentration, comprising:

[0090] S510, the garbage concentration of each bounding box in the cleaning area is accumulated and then exponentially normalized to obtain the cleaning difficulty of the cleaning area;

[0091] S520, a range value of the bristle angle is obtained, and a first functional relationship between the cleaning difficulty and the bristle angle is established within the range value; and the bristle angle is adjusted based on the cleaning difficulty within the range value;

[0092] Specifically, the angle of the bristles of the cleaning brush is adjusted according to the cleaning difficulty of the garbage. For example, for the cleaning area where there is a large amount of garbage or a large amount of garbage that is difficult to clean, the bristle angle is increased, so that the bristles can better pick up the garbage and sweep it into the garbage can; for the cleaning area where there is small particles and other garbage with small cleaning difficulty, the bristle angle is reduced, the contact area of the bristles with the ground is increased, and the cleaning effect is improved.

[0093] The relationship formula between the bristle angle θ and the cleaning difficulty S is:

[0094] ;

[0095] wherein, and are the maximum and minimum values of the bristle angle, is the reference step length of the bristle angle, is the threshold value of the cleaning difficulty, is the cleaning difficulty, and k is an adjustment coefficient, 0

[0096] S530, the rotation speed of the cleaning brush, the driving speed of the cleaning vehicle, the reference speed of the cleaning vehicle and the basic rotation speed of the cleaning brush are obtained, and the rotation speed of the cleaning brush is determined in combination with the cleaning difficulty of the garbage.

[0097] Specifically, the rotation speed of the cleaning brush is adjusted in real time according to the cleaning difficulty of the garbage and the driving speed of the cleaning vehicle. In the garbage dense area or when the cleaning vehicle drives slowly, the rotation speed of the cleaning brush is increased to increase the cleaning strength; in the garbage sparse area or when the cleaning vehicle drives fast, the rotation speed of the cleaning brush is reduced to save energy.

[0098] The relationship formula of the rotation speed n of the cleaning brush, the cleaning difficulty S and the driving speed v of the cleaning vehicle is:

[0099]

[0100] wherein, n is the rotation speed of the cleaning brush, S is the cleaning difficulty of the garbage, v is the driving speed of the cleaning vehicle, v0 is the reference speed of the cleaning vehicle, n0 is the basic rotation speed of the cleaning brush, and a and b are weight coefficients of the cleaning difficulty and the driving speed of the cleaning vehicle, respectively.

[0101] During the driving of the sweeper, the control strategy of the sweeping brush is dynamically adjusted according to the real-time change of the garbage cleaning difficulty. When the sweeper is about to reach the area with large cleaning difficulty, the rotating speed, angle and position of the sweeping brush are adjusted in advance to ensure that the sweeping brush can clean the garbage in time and effectively; when the sweeper drives away from the garbage area, the working state of the sweeping brush is adjusted correspondingly to reduce the energy consumption.

[0102] Reference Figure 2 The embodiment of the present application also provides a sweeping brush control system of a sweeper based on image recognition, which comprises:

[0103] at least one processor;

[0104] at least one memory for storing at least one program;

[0105] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.

[0106] The contents in the above method embodiments are all applicable to the present embodiment, the function specifically implemented by the present embodiment is the same as that of the above method embodiments, and the beneficial effects achieved by the present embodiment are also the same as those achieved by the above method embodiments, which will not be repeated here.

[0107] In addition, the embodiment of the present application also discloses a computer readable storage medium, which stores computer executable instructions, and the computer executable instructions are used for executing the above method.

[0108] In addition, the embodiment of the present application also discloses a computer program product or a computer program, which is stored in a computer readable storage medium. The processor of the computer equipment can read the computer program from the computer readable storage medium, and the processor executes the computer program, so that the computer equipment executes the above method. Similarly, the contents in the above method embodiments are all applicable to the present storage medium embodiment, the function specifically implemented by the present storage medium embodiment is the same as that of the above method embodiments, and the beneficial effects achieved by the present storage medium embodiment are also the same as those achieved by the above method embodiments.

[0109] Those of ordinary skill in the art will realize and understand that all or some of the systems in the methods disclosed above can be implemented as software, firmware, hardware, and appropriate combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software can be distributed on computer readable media, which can include computer storage media (or non-transitory media) and communication media (or transitory media). As is well known to those of ordinary skill in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer. Further, it is common in the art for a person of ordinary skill to understand that communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.

[0110] While the description of the present disclosure has been quite specific and has been described with reference to several described embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, but instead it is intended to cover the intended scope of the present disclosure as broadly as the appended claims provide such scope, in view of the prior art, and with the full scope of equivalents to which such claims are entitled. Furthermore, the foregoing description of the present disclosure has been set forth in the preceding description with reference to embodiments that are presently contemplated as examples of the present disclosure. It is intended that changes in details, including improvements and substitutions, can be made by those skilled in the art, that the scope of the present disclosure is not limited to the specific examples described, and that the present disclosure is to be afforded the broadest interpretation of the appended claims, as interpreted in the light of the prior art.

Claims

1. An image recognition-based sweeper sweeper brush control method, characterized by, The method comprises the following steps: An image of a road surface is acquired, and a target detection model is used to extract a bounding box of each type of garbage in the road image; A garbage type probability distribution of pixels in each bounding box is determined, and a cross degree of each type of garbage in the bounding box is calculated based on the garbage type probability distribution and the pixel position of each type of garbage; A pollution weight of the bounding box is established based on the garbage type of each pixel in the bounding box, and a garbage concentration of each bounding box is determined based on the cross degree and the pollution weight of each type of garbage in each bounding box; Based on the garbage concentration and the position distribution of each bounding box, clustering analysis is performed on each bounding box to obtain a cleaning area; Based on the garbage concentration of each bounding box in the cleaning area, a cleaning difficulty of the cleaning area is determined, and the angle and rotating speed of the cleaning brush in each cleaning area are controlled according to the cleaning difficulty; The determination of the garbage type probability distribution of pixels in each bounding box and the calculation of the cross degree of garbage in the bounding box based on the garbage type probability distribution and the pixel position of each type of garbage comprise: The ratio of the frequency of each pixel in the bounding box belonging to the target type of garbage to the total number of pixels is accumulated to obtain the probability of the target type of garbage in the bounding box; The pixel position difference of the target type of garbage in two bounding boxes is calculated to obtain the pixel difference of the target type of garbage in the two bounding boxes; The probability and pixel difference of each target type of garbage in the two bounding boxes are multiplied and accumulated to obtain the cross degree of the target type of garbage in the two bounding boxes, and the cross degrees of each target type of garbage in the bounding box are accumulated to obtain the cross degrees of each type of garbage in the bounding box; The establishment of the pollution weight of the bounding box based on the garbage type of each pixel in the bounding box and the determination of the garbage concentration of each bounding box based on the cross degree and the pollution weight of each type of garbage in each bounding box comprise: A basic weight coefficient is set according to the garbage type, the garbage concentration of each type of garbage in the bounding box is multiplied by the corresponding basic weight coefficient to obtain the pollution weight of each type of garbage; The pollution weights of each type of garbage in the bounding box are weighted and summed with the cross degrees to obtain a comprehensive pollution index of the bounding box; Based on the comprehensive pollution index of the bounding box and the area of the bounding box, a garbage concentration value per unit area is calculated as the garbage concentration of the bounding box; The clustering analysis of each bounding box based on the garbage concentration and the position distribution of each bounding box to obtain a cleaning area comprises: S410, a cleaning radius is set according to the coverage area of the cleaning brush and the average value of the garbage concentration of each bounding box, a minimum point number is set according to the minimum garbage number in the coverage area of the road image, and a clustering center is determined based on the garbage concentration of each bounding box; S420, the cleaning radius is used as the clustering boundary of the clustering center, and the adjacent bounding boxes of the clustering center are clustered within the minimum point number to form a cleaning area; S430, the cleaning area is equally divided into multiple sub-areas, and a local garbage concentration of each sub-area is determined based on the area proportion and the garbage concentration of each bounding box in the sub-area. S440, respectively determine the distance of each sub-region and the cluster center, select the sub-region with the maximum product of local garbage concentration and distance as the new cluster center, determine whether the distance change of the cluster center is lower than the distance threshold, if yes, execute S450; if no, execute S420; S450, gradient calculation is performed on the garbage concentration of each bounding box in the cleaning area, a plurality of bounding boxes with larger gradient are taken as reference boxes, an external curve containing the reference boxes is generated, and a cleaning area containing the external curve is generated with the cluster center as the center and within the cleaning radius and the minimum point number.

2. The method of claim 1, wherein, The cluster boundary with the cleaning radius as the cluster center clusters the adjacent bounding boxes of the cluster center within the minimum point number to form a cleaning area, including: Filtering the adjacent bounding boxes according to the cluster center and the cleaning radius, ensuring that the number of bounding boxes in each cluster is not less than the minimum point number, and forming a cleaning unit; If there is an overlapping bounding box on the edge of the cleaning unit, the bounding box is divided into a cleaning unit with a larger pixel ratio, the cleaning unit is adjusted in position based on the bounding box in the cleaning unit, and the adjusted cleaning unit is taken as the cleaning area.

3. The method of claim 1, wherein, The cleaning difficulty of the cleaning area is determined based on the garbage concentration of each bounding box in the cleaning area, and the angle and rotating speed of the cleaning brush in each cleaning area are controlled according to the cleaning difficulty, including: The garbage concentration of each bounding box in the cleaning area is accumulated and then subjected to exponential normalization processing to obtain the cleaning difficulty of the cleaning area; A range value of the bristle angle is obtained, a first functional relationship between the cleaning difficulty and the bristle angle is established within the range value, and the bristle angle is adjusted within the range value based on the cleaning difficulty; The rotating speed of the cleaning brush, the driving speed of the cleaning vehicle, the reference speed of the cleaning vehicle and the basic rotating speed of the cleaning brush are obtained, and the rotating speed of the cleaning brush is determined in combination with the cleaning difficulty of the garbage.

4. An image recognition-based sweeper brush control system of a sweeper, characterized by, Including: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method of any one of claims 1 to 3.

5. A computer readable storage medium having stored therein a program which is executable by a processor, characterized in that, The program executable by the processor when executed by the processor is used to execute the method of any one of claims 1 to 3.

Images