Unmanned sanitation vehicle sweeping distance measurement method based on image recognition

Abstract

The invention relates to the technical field of unmanned driving, and discloses an unmanned sanitation vehicle sweeping distance measurement method based on image recognition. An unmanned sanitation vehicle monitoring module, a small target object amplification module, an unmanned sweeping data collection module, an unmanned sweeping data analysis module, an unmanned sanitation vehicle data evaluation module, an unmanned sanitation vehicle data optimization module and an unmanned sanitation vehicle control module are established. The unmanned sanitation vehicle monitoring module collects image data, the collected small target object image is amplified in the small target object amplification module, the unmanned cleaning data collection module collects data, the unmanned cleaning data analysis module analyzes the collected data, and the unmanned sanitation vehicle monitoring module sends the analyzed data to the unmanned sanitation vehicle monitoring module. The unmanned sanitation vehicle data evaluation module evaluates the performance of a sanitation vehicle, the unmanned sanitation vehicle data optimization module optimizes parameters, and the unmanned sanitation vehicle control module issues an adjustment instruction according to the optimized parameters.

Description

Technical Field

[0001] The present invention relates to the field of unmanned driving technology, and in particular to a cleaning distance measurement method for an unmanned driving sanitation vehicle based on image recognition. Background Art

[0002] In existing technologies, unmanned sanitation vehicles face many challenges when identifying small target objects. Due to the inconsistent distance between the target object and the vehicle camera, the target object appears at different scales in the collected image. For small target objects at a long distance, the pixel area they occupy on the image is small, and the information is insufficient, making recognition difficult. Most of the existing target detection methods are optimized for large and medium-sized target objects of a certain size, while for small target objects with small scale and unclear features, the recognition accuracy is difficult to achieve ideal results. In addition, small target objects usually lack rich texture and color features and are easily mistaken for background noise. Therefore, how to improve the recognition accuracy of small target objects of unmanned sanitation vehicles in complex environments has become an urgent problem to be solved in the current technological development. Summary of the invention

[0003] 1. Technical issues to be resolved

[0004] In view of the shortcomings of the prior art, the present invention provides a cleaning and ranging method for an unmanned sanitation vehicle based on image recognition, which has the advantage of high recognition accuracy of small target objects in complex environments, and solves the problem of low recognition accuracy of small target objects by unmanned sanitation vehicles in complex environments.

[0005] (II) Technical solution

[0006] To achieve the above object, the present invention provides the following technical solution: a method for measuring the distance between an unmanned sanitation vehicle and a cleaning machine based on image recognition, comprising the following steps:

[0007] Step 1: On the basis of the unmanned sanitation vehicle control system, establish an unmanned sanitation vehicle monitoring module, a small target object magnification module, an unmanned cleaning data collection module, an unmanned cleaning data analysis module, an unmanned sanitation vehicle data evaluation module, an unmanned sanitation vehicle data optimization module and an unmanned sanitation vehicle control module;

[0008] Step 2: The unmanned sanitation vehicle monitoring module monitors the environment in real time and collects image data through sensors, high-definition cameras and lidar;

[0009] Step 3: In the small target object magnification module, the collected small target object image is magnified, and the image enhancement technology is used to perform deep recognition of the small target object;

[0010] Step 4: The unmanned cleaning data collection module collects data from the unmanned sanitation vehicle monitoring module and the small target object magnification module;

[0011] Step 5: The unmanned cleaning data analysis module analyzes the collected data, including the recognition and ranging of small target objects;

[0012] Step 6: The unmanned sanitation vehicle data evaluation module evaluates the performance of the sanitation vehicle through analysis results, and can make timely adjustments or eliminate problems when poor status or failures are detected;

[0013] Step 7: The unmanned sanitation vehicle data optimization module optimizes the control parameters of the unmanned sanitation vehicle according to the analysis results;

[0014] Step 8. The unmanned sanitation vehicle control module issues adjustment instructions according to the optimized parameters to control the sanitation vehicle to perform cleaning.

[0015] Preferably, the unmanned driving cleaning data collection module includes an image and perception unit, a sensor and positioning unit, and a vehicle status and control unit. The image and perception unit collects environmental images and enlarged photos of small target objects, and obtains image data and environmental perception data through high-precision cameras and weather stations. The sensor and positioning unit obtains three-dimensional information and distance measurement of the vehicle's surroundings, and obtains positioning data through sensors and lidars. The vehicle status and control unit provides operating parameters of the vehicle's speed, acceleration and direction, and obtains vehicle status data through the unmanned driving sanitation vehicle control system. The image and perception unit, sensor and positioning unit, and vehicle status and control unit number the internal data of the unit and connect to the unmanned driving cleaning data analysis module through the network.

[0016] Preferably, the unmanned driving cleaning data analysis module includes a small target recognition unit, a small target ranging unit and a method feasibility detection unit.

[0017] Preferably, the image data is numbered f 1 、f 2 、f 3 ,…,f n , the environmental perception data number is H 1 , H 2 , H 3 ,…,H n .

[0018] Preferably, the positioning data includes the focal length of the camera, the baseline distance between the two cameras and the parallax of the objects in the two images, and the focal length of the camera, the baseline distance between the two cameras and the parallax of the objects in the two images are numbered k, l, and d respectively.

[0019] Preferably, the vehicle status data number is G 1 , G 2 , G 3 , …, G m .

[0020] Preferably, the small target recognition unit calculates the probability Pz of small target object recognition based on the image data and the environmental perception data, and the calculation formula is:

[0021]

[0022] In the formula, Pz represents the probability of small target object recognition, f 1 、f 2 、f 3 ,…,f n Represents image data, H 1 , H 2 , H 3 ,…,H n represents environmental perception data, n represents the number of image data points, and f i represents the i-th image data, f i (t i ) represents the image feature vector corresponding to the i-th image data, which is extracted by the CNN technology model. i represents the i-th environmental perception data, w i The weight of the environmental impact image in the i-th environmental perception data is extracted through the CNN technology model, where a represents the wind speed impact factor in the environment and e is a constant;

[0023] The probability P of small target object recognition is used for recognition of small target objects.

[0024] Preferably, the small target distance measuring unit calculates the distance Dr of the small target object according to the positioning data, and the calculation formula is:

[0025]

[0026] In the formula, Dr1 represents the distance of the small target object, k represents the focal length of the camera, l represents the baseline distance between the two cameras, and d represents the parallax of the object in the two images;

[0027] The distance Dr of the small target object is used for measuring the distance of the small target object.

[0028] Preferably, the method feasibility detection unit calculates the evaluation value Dr of the effectiveness of the sanitation vehicle system according to the vehicle status data, and the calculation formula is:

[0029]

[0030] In the formula, Dr represents the evaluation value of the effectiveness of the sanitation vehicle system, Z j represents the accuracy of identifying the jth small target object after using this ranging method, X j represents the recall rate of the i-th small target object recognition after using this ranging method, G 1 , G 2 , G 3 , …, G m Represents vehicle status data, G j represents the evaluation value of the j-th vehicle status data after using this ranging method, μ j , b j , α j They represent the weight coefficients corresponding to the accuracy, recall rate and evaluation value of the vehicle status data of the jth small target object recognition, and satisfy μ j +b j +α j =1, m is the number of evaluations or data points;

[0031] The evaluation value Dr of the effectiveness of the sanitation vehicle system is used to evaluate the feasibility of this distance measurement method.

[0032] Preferably, the unmanned sanitation vehicle data evaluation module monitors the performance of the sanitation vehicle through the evaluation value Dr of the effectiveness of the sanitation vehicle system. When it is detected that the evaluation value Dr of the effectiveness of the sanitation vehicle system is lower than a preset threshold, the module triggers an automatic adjustment mechanism to fine-tune the parameters of the sanitation vehicle;

[0033] The unmanned sanitation vehicle data optimization module receives feedback information from the unmanned sanitation vehicle data evaluation module, as well as the probability P of small target object recognition and the distance Dr of the small target object from the data analysis module, and optimizes the control parameters of the sanitation vehicle according to the probability P of small target object recognition and the distance Dr of the small target object.

[0034] Compared with the prior art, the present invention provides a cleaning distance measurement method for an unmanned sanitation vehicle based on image recognition, which has the following beneficial effects:

[0035] 1. The present invention receives feedback information from the unmanned sanitation vehicle data evaluation module and the probability P of small target object recognition and the distance Dr of the small target object from the data analysis module through the unmanned sanitation vehicle data optimization module, and optimizes the control parameters of the sanitation vehicle and the cleaning speed according to the probability P of small target object recognition and the distance Dr of the small target object. When there are many small target objects recognized and the distance is close, it is necessary to reduce the cleaning speed to ensure the cleaning quality, optimize the cleaning path, and adjust the cleaning path to give priority to cleaning the small target object area with high recognition probability and close distance, optimize the suction or cleaning force, and solve the problem of low recognition accuracy of small target objects by unmanned sanitation vehicles in complex environments according to the probability P of small target object recognition and the distance Dr of the small target object. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of the structure of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] See also Figure 1 , a cleaning distance measurement method for an unmanned sanitation vehicle based on image recognition, comprising the following steps:

[0039] Step 1: On the basis of the unmanned sanitation vehicle control system, establish an unmanned sanitation vehicle monitoring module, a small target object magnification module, an unmanned cleaning data collection module, an unmanned cleaning data analysis module, an unmanned sanitation vehicle data evaluation module, an unmanned sanitation vehicle data optimization module and an unmanned sanitation vehicle control module;

[0040] Step 2: The unmanned sanitation vehicle monitoring module monitors the environment in real time and collects image data through sensors, high-definition cameras and lidar;

[0041] Step 3: In the small target object magnification module, the collected small target object image is magnified, and the image enhancement technology is used to perform deep recognition of the small target object;

[0042] Step 4: The unmanned cleaning data collection module collects data from the unmanned sanitation vehicle monitoring module and the small target object magnification module;

[0043] Step 5: The unmanned cleaning data analysis module analyzes the collected data, including the recognition and ranging of small target objects;

[0044] Step 6: The unmanned sanitation vehicle data evaluation module evaluates the performance of the sanitation vehicle through analysis results, and can make timely adjustments or eliminate problems when poor status or failures are detected;

[0045] Step 7: The unmanned sanitation vehicle data optimization module optimizes the control parameters of the unmanned sanitation vehicle according to the analysis results;

[0046] Step 8. The unmanned sanitation vehicle control module issues adjustment instructions according to the optimized parameters to control the sanitation vehicle to perform cleaning.

[0047] The unmanned cleaning data collection module includes an image and perception unit, a sensor and positioning unit, and a vehicle status and control unit. The image and perception unit collects environmental images and enlarged photos of small target objects, and obtains image data and environmental perception data through high-precision cameras and weather stations. The sensor and positioning unit obtains three-dimensional information and distance measurement of the vehicle's surroundings, and obtains positioning data through sensors and lidars. The vehicle status and control unit provides the vehicle's operating parameters of speed, acceleration and direction, and obtains vehicle status data through the unmanned sanitation vehicle control system. The image and perception unit, sensor and positioning unit, and vehicle status and control unit number the internal data of the unit and connect to the unmanned cleaning data analysis module through the network.

[0048] The unmanned driving cleaning data analysis module includes a small target recognition unit, a small target ranging unit and a method feasibility detection unit.

[0049] The image data number is f 1 、f 2 、f 3 ,…,f n , the environmental perception data number is H 1 , H 2 , H 3 ,…,H n .

[0050] The positioning data includes the focal length of the camera, the baseline distance between the two cameras, and the parallax of the objects in the two images. The focal length of the camera, the baseline distance between the two cameras, and the parallax of the objects in the two images are numbered k, l, and d, respectively.

[0051] The vehicle status data number is G 1 , G 2 , G 3 , …, G m .

[0052] The small target recognition unit calculates the probability Pz of small target object recognition based on image data and environmental perception data. The calculation formula is:

[0053]

[0054] In the formula, Pz represents the probability of small target object recognition, f 1 、f 2 、f 3 ,…,f n Represents image data, H 1 , H 2 , H 3 ,…,H n represents environmental perception data, n represents the number of image data points, and f i represents the i-th image data, f i (t i ) represents the image feature vector corresponding to the i-th image data, which is extracted by the CNN technology model. i represents the i-th environmental perception data, w i The weight of the environmental impact image in the i-th environmental perception data is extracted through the CNN technology model, where a represents the wind speed impact factor in the environment and e is a constant;

[0055] The advantages are: by calculating the probability Pz of small target object recognition, the probability P of small target object recognition is used for the recognition of small target objects. The formula comprehensively considers image data and perception data in complex environments, and calculates the probability Pz of small target object recognition by weighted averaging the above data. In actual implementation application, this value can be used to evaluate the small target object recognition ability of unmanned sanitation vehicles in complex environments, thereby improving the accuracy of unmanned sanitation vehicles in recognizing small target objects in complex environments, so as to improve the recognition efficiency of the sanitation vehicle control system.

[0056] The small target ranging unit calculates the distance Dr of the small target object based on the positioning data. The calculation formula is:

[0057]

[0058] In the formula, Dr1 represents the distance of the small target object, k represents the focal length of the camera, l represents the baseline distance between the two cameras, and d represents the parallax of the object in the two images;

[0059] The advantages are: by calculating the distance Dr of the small target object, the distance Dr of the small target object is used for ranging of the small target object. This formula is a ranging method based on stereo vision, and the distance is estimated by calculating the parallax of the object in the image captured by two cameras. In actual implementation application, this value can be used by unmanned sanitation vehicles to accurately measure the distance of small target objects, thereby improving the operating efficiency of the sanitation vehicle and the accuracy of cleaning the position of small target objects.

[0060] The feasibility detection unit of the method calculates the evaluation value Dr of the effectiveness of the sanitation vehicle system according to the vehicle status data, and the calculation formula is:

[0061]

[0062] In the formula, Dr represents the evaluation value of the effectiveness of the sanitation vehicle system, Z j represents the accuracy of identifying the jth small target object after using this ranging method, X j represents the recall rate of the i-th small target object recognition after using this ranging method, G 1 , G 2 , G 3 , …, G m Represents vehicle status data, G j represents the evaluation value of the j-th vehicle state data (such as speed and acceleration) after using this ranging method (the evaluation value of the vehicle state data is derived from the vehicle state data), μ j 、b j , α j They represent the weight coefficients corresponding to the accuracy, recall rate and evaluation value of vehicle status data (such as speed and acceleration) of the jth small target object recognition, and satisfy μ j +b j +α j =1, m is the number of evaluations or data points;

[0063] The advantages are: by calculating the evaluation value Dr of the sanitation vehicle system effectiveness, the evaluation value Dr of the sanitation vehicle system effectiveness is used to evaluate the feasibility of the present ranging method. The calculation formula comprehensively considers the small target object recognition accuracy and recall rate after using the present ranging method, as well as the vehicle status data, to calculate the evaluation value Dr of the sanitation vehicle system effectiveness. In actual implementation application, this value can be used to evaluate the feasibility of the present ranging method, ensure the performance and reliability of unmanned sanitation vehicles in complex environments, and thus ensure the normal progress of cleaning work.

[0064] The driverless sanitation vehicle data evaluation module monitors the performance of the sanitation vehicle through the evaluation value Dr of the sanitation vehicle system effectiveness. When it is detected that the evaluation value Dr of the sanitation vehicle system effectiveness is lower than the preset threshold, the module will trigger the automatic adjustment mechanism to fine-tune the parameters of the sanitation vehicle, such as adjusting the focal length of the camera and the sensitivity of the sensor. If the performance still does not meet the standard after adjustment, the module will further diagnose the obstacles or faults that may occur, and try to eliminate them by restarting the system and cleaning the sensors.

[0065] The unmanned sanitation vehicle data optimization module receives feedback information from the unmanned sanitation vehicle data evaluation module, as well as the probability P of small target object recognition and the distance Dr of small target objects from the data analysis module. Combined with the small target object recognition probability and distance data of the data analysis module, the sanitation vehicle can more accurately identify and locate small target objects. This accuracy provides a reliable foundation for subsequent cleaning operations, helps to improve the cleaning quality, ensure the cleanliness of the cleaning area, and optimize the control parameters of the sanitation vehicle according to the probability P of small target object recognition and the distance Dr of small target objects, such as: Cleaning speed: When there are many small target objects recognized and the distance is close, the cleaning speed needs to be reduced to ensure the cleaning quality and avoid cleaning errors caused by excessive speed. This targeted adjustment enables the sanitation vehicle to achieve the best cleaning effect with minimal energy consumption, further improving the cleaning efficiency; cleaning path: adjust the cleaning path to give priority to cleaning the small target object area with high recognition probability and close distance; suction or cleaning force: adjust the suction or cleaning force according to the type and distance of the small target object to improve the cleaning efficiency. The sanitation vehicle can deal with these key areas more intensively. This targeted cleaning strategy can effectively avoid the situation where small target objects are missed or not cleaned thoroughly, and ensure that the cleaning quality meets the expected standards; this real-time adjustment mechanism enables the sanitation vehicle to quickly adapt to different cleaning environments and task requirements, enhancing the adaptability and flexibility of the system;

[0066] The unmanned sanitation vehicle control module issues adjustment instructions based on the optimized parameters to control the sanitation vehicle to clean. Its working process is as follows:

[0067] S1. Receive optimization parameters from a data optimization module;

[0068] S2, generating specific control instructions based on these parameters, such as adjusting the speed, direction, and suction of the sanitation vehicle;

[0069] S3, transmitting control instructions to the actuators of the sanitation vehicle, such as the motor and hydraulic system, to achieve precise control of the sanitation vehicle;

[0070] S4. Monitor the performance of the sanitation vehicle to ensure that it performs cleaning operations according to the optimized parameters;

[0071] Through the collaborative work of the above modules, the unmanned sanitation vehicle can complete the cleaning task efficiently and accurately in complex environments. At the same time, it has the ability to self-adjust and optimize, improving the reliability and adaptability of the system. By reasonably adjusting the operating parameters of the sanitation vehicle, it can effectively avoid the equipment from running under unnecessary high load conditions, thereby reducing equipment wear, helping to extend the service life of the sanitation vehicle, reduce equipment maintenance and replacement costs, and further improve economic benefits.

[0072] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for measuring the distance between an unmanned sanitation vehicle and a cleaning machine based on image recognition, characterized in that: The following steps are involved: Step 1: On the basis of the unmanned sanitation vehicle control system, establish an unmanned sanitation vehicle monitoring module, a small target object magnification module, an unmanned cleaning data collection module, an unmanned cleaning data analysis module, an unmanned sanitation vehicle data evaluation module, an unmanned sanitation vehicle data optimization module and an unmanned sanitation vehicle control module; Step 2: The unmanned sanitation vehicle monitoring module monitors the environment in real time and collects image data through sensors, high-definition cameras and lidar; Step 3: In the small target object magnification module, the collected small target object image is magnified, and the image enhancement technology is used to perform deep recognition of the small target object; Step 4: The unmanned cleaning data collection module collects data from the unmanned sanitation vehicle monitoring module and the small target object magnification module; Step 5: The unmanned cleaning data analysis module analyzes the collected data, including the recognition and ranging of small target objects; Step 6: The unmanned sanitation vehicle data evaluation module evaluates the performance of the sanitation vehicle through analysis results, and can make timely adjustments or eliminate problems when poor status or failures are detected; Step 7: The unmanned sanitation vehicle data optimization module optimizes the control parameters of the unmanned sanitation vehicle according to the analysis results; Step 8. The unmanned sanitation vehicle control module issues adjustment instructions according to the optimized parameters to control the sanitation vehicle to perform cleaning.

2. According to claim 1, a method for measuring distance of an unmanned sanitation vehicle based on image recognition, characterized in that: The unmanned driving cleaning data collection module includes an image and perception unit, a sensor and positioning unit, and a vehicle status and control unit. The image and perception unit collects environmental images and enlarged photos of small target objects, and obtains image data and environmental perception data through high-precision cameras and weather stations. The sensor and positioning unit obtains three-dimensional information and distance measurement of the vehicle's surrounding environment, and obtains positioning data through sensors and lidars. The vehicle status and control unit provides operating parameters of the vehicle's speed, acceleration and direction, and obtains vehicle status data through the unmanned driving sanitation vehicle control system. The image and perception unit, sensor and positioning unit, and vehicle status and control unit number the internal data of the unit and connect to the unmanned driving cleaning data analysis module through the network.

3. The method for measuring distance by cleaning an unmanned sanitation vehicle based on image recognition according to claim 2 is characterized in that: The unmanned driving cleaning data analysis module includes a small target recognition unit, a small target ranging unit and a method feasibility detection unit.

4. The method for measuring distance by an unmanned sanitation vehicle based on image recognition according to claim 2 is characterized in that: The image data are numbered f1, f2, f3, ..., f n , the environmental perception data are numbered H1, H2, H3, ..., H n .

5. The method for measuring distance of an unmanned sanitation vehicle based on image recognition according to claim 2 is characterized in that: The positioning data includes the focal length of the camera, the baseline distance between the two cameras, and the parallax of the objects in the two images, and the focal length of the camera, the baseline distance between the two cameras, and the parallax of the objects in the two images are numbered k, l, and d respectively.

6. The method for measuring distance of an unmanned sanitation vehicle based on image recognition according to claim 2, characterized in that: The vehicle status data are numbered G1, G2, G3, ..., G m .

7. The method for measuring distance by cleaning an unmanned sanitation vehicle based on image recognition according to claim 4 is characterized in that: The small target recognition unit calculates the probability Pz of small target object recognition based on the image data and the environmental perception data, and the calculation formula is: In the formula, Pz represents the probability of small target object recognition, f1, f2, f3, ..., f n Represents image data, H1, H2, H3, ..., H n represents environmental perception data, n represents the number of image data points, and f i represents the i-th image data, f i (t i ) represents the image feature vector corresponding to the i-th image data, which is extracted by the CNN technology model. i represents the i-th environmental perception data, w i The weight of the environmental impact image in the i-th environmental perception data is extracted through the CNN technology model, where a represents the wind speed impact factor in the environment and e is a constant; The probability P of small target object recognition is used for recognition of small target objects.

8. The method for measuring distance by cleaning an unmanned sanitation vehicle based on image recognition according to claim 5, characterized in that: The small target distance measuring unit calculates the distance Dr of the small target object according to the positioning data, and the calculation formula is: In the formula, Dr1 represents the distance of the small target object, k represents the focal length of the camera, l represents the baseline distance between the two cameras, and d represents the parallax of the object in the two images; The distance Dr of the small target object is used for measuring the distance of the small target object.

9. The method for measuring distance by cleaning an unmanned sanitation vehicle based on image recognition according to claim 6, characterized in that: The feasibility detection unit of the method calculates the evaluation value Dr of the effectiveness of the sanitation vehicle system according to the vehicle status data, and the calculation formula is: In the formula, Dr represents the evaluation value of the effectiveness of the sanitation vehicle system, Z j represents the accuracy of identifying the jth small target object after using this ranging method, X j represents the recall rate of the i-th small target object recognition after using this ranging method, G1, G2, G3, ..., G m Represents vehicle status data, G j represents the evaluation value of the j-th vehicle status data after using this ranging method, μ j 、b j , α j They represent the weight coefficients corresponding to the accuracy, recall rate and evaluation value of the vehicle status data of the jth small target object recognition, and satisfy μ j +b j +α j =1, m is the number of evaluations or data points; The evaluation value Dr of the effectiveness of the sanitation vehicle system is used to evaluate the feasibility of this distance measurement method.

10. The method for cleaning and measuring distance of an unmanned sanitation vehicle based on image recognition according to claim 9, characterized in that: The unmanned sanitation vehicle data evaluation module monitors the performance of the sanitation vehicle through the evaluation value Dr of the sanitation vehicle system effectiveness. When it is detected that the evaluation value Dr of the sanitation vehicle system effectiveness is lower than the preset threshold, the module will trigger the automatic adjustment mechanism to fine-tune the parameters of the sanitation vehicle; The unmanned sanitation vehicle data optimization module receives feedback information from the unmanned sanitation vehicle data evaluation module, as well as the probability P of small target object recognition and the distance Dr of the small target object from the data analysis module, and optimizes the control parameters of the sanitation vehicle according to the probability P of small target object recognition and the distance Dr of the small target object.

Images