Intelligent subway track cleaning device and method based on embedded machine vision

Abstract

The present invention discloses an intelligent subway track cleaning device and method based on embedded machine vision. The device includes: an electric-controlled cleaning vehicle, a machine vision system, a track operation platform and a power supply system. The electric-controlled cleaning vehicle moves according to the garbage location coordinates provided by the machine vision system. The track operation platform includes: a transverse slide rail, a cleaning and dust collection component and a garbage collection box. The transverse slide rail is arranged above the electric-controlled cleaning vehicle; the cleaning and dust collection component is controlled to start and stop by a controller, and the cleaning and dust collection component is driven to move laterally on the transverse slide rail according to the garbage location coordinates; the garbage collection box is arranged on the electric-controlled cleaning vehicle, and plastic bags, dust and other garbage are cleaned and sucked into the garbage collection box for storage by a vacuum cleaner. The cleaning device of the present invention can accurately identify garbage on the subway track, and through the cooperation of the electric-controlled cleaning vehicle and the track operation platform, the scope of garbage cleaning is greatly expanded.

Description

Technical Field

[0001] The present invention relates to the technical field of subway track cleaning, and in particular to an intelligent subway track cleaning device and method based on embedded machine vision. Background Art

[0002] As a fast, high-capacity, energy-efficient, and environmentally friendly means of urban passenger transportation, subways are increasingly popular in cities both domestically and internationally. However, many subway stations are plagued by small advertising waste. Trash thrown onto the platform is swept into the gaps between the doors by wind, while passengers throw trash between the platform and the screen door, which then falls directly onto the subway tracks, posing a safety hazard. Surveys show that waste paper and food bags are the most common waste. Every morning, passengers throw their empty soy milk cups or food bags onto the platform or between the doors of subway cars. Subway stations generally use three methods for waste disposal: daily, weekly, and monthly. Daily and weekly collection primarily involves surface-level waste removal, while monthly cleaning also involves cleaning sludge and dust from gaps, a significant workload. Furthermore, there have been incidents of track fires caused by accumulated trash at subway stations, resulting in line closures and injuries. Therefore, regular cleaning of subway tracks and platforms is essential to ensure safe operation and a clean platform environment.

[0003] Currently, subway waste disposal relies primarily on manpower and traditional cleaning machines, such as sweepers and cleaning vehicles. These traditional cleaning machines not only require manual operation to clean trash from subway tracks and platforms, but their large size also makes them difficult to maneuver within the subway tracks. Most can only shuttle along the subway lines to collect trash in a single pass. To ensure their passability, their cleaning range is limited to the narrow area inside and outside the subway tracks. Specific areas and those near the edges of the passages still require manual cleaning. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent subway track cleaning device and method based on embedded machine vision, which can accurately locate garbage on the inside, outside and outside of the subway track near the edge of the channel, and accurately clean the garbage at the located position.

[0005] The technical solution of the present invention is:

[0006] An intelligent subway track cleaning device based on embedded machine vision includes: an electric-controlled cleaning vehicle equipped with a drive assembly capable of moving forward and backward along the subway track; a track operating platform including: a transverse slide rail, the sliding direction of which is toward both sides of the subway track, the transverse slide rail being arranged above the electric-controlled cleaning vehicle through a bracket; a cleaning and dust collection assembly including: a garbage collection box, which is slidably arranged on the transverse slide rail through a sliding wheel assembly; a plurality of vacuum cleaners arranged at the bottom of the garbage collection box, and the dust collection pipes of the vacuum cleaners are arranged vertically, the output ends of the dust collection pipes being connected to the garbage collection box, for directly cleaning and sucking garbage such as plastic bags and dust into the garbage collection box for storage; a machine vision system arranged on the electric-controlled cleaning vehicle, the machine vision system serving as a carrier for running a main program including: : A controller is connected to the drive assembly of the electric-controlled cleaning vehicle and is connected and communicated with the control module of the vacuum cleaner; an image recognition system includes: an image acquisition module and an image processing module, the image acquisition module is connected to the image processing module, and the image processing module is connected and communicated with the controller; an autonomous route planning system is connected and communicated with the controller, for receiving the image data processed by the image processing module and calculating the movement path of the electric-controlled cleaning vehicle to the coordinates of the garbage location based on the processed image data, using GPS positioning for autonomous cruising, using GPS as the source of positioning coordinates, and using the Baidu map API embedded in the ground station as the data source. Since the GPS heading refresh frequency is low, the HMC5883L digital compass is selected as the source of heading data. The problem of weak underground signals can be solved by calculating the relative position through the inertial navigation function of the GPS module or by using a signal repeater.

[0007] Furthermore, the system also includes a power supply system for the electrically controlled cleaning vehicle, the machine vision system, and the track-based operating platform. This power supply system utilizes a third-rail power supply method, with a current collector provided on the track-based operating platform and connected to the third rail. Powering the system through the third rail not only reduces the size of the cleaning device and ensures its flexibility, but also ensures stable operation and power quality, enabling the cleaning device to operate within the relatively narrow subway track environment.

[0008] Furthermore, the machine vision system leverages the existing OPENCV library to build an embedded system hardware platform. It employs a recognition algorithm that combines image enhancement, adaptive binarization, Canny edge detection, and convolutional neural networks to identify trash within subway tracks. This system can adapt to the complex environment of subway tracks, accurately and efficiently identifying small and light trash within them, and performing location and path planning. This allows the track operating platform and the electronically controlled cleaning vehicle to move together, enabling unmanned and autonomous operation of the cleaning device.

[0009] Furthermore, a screw transmission assembly is arranged on the transverse slide rail, and the screw transmission assembly has a transmission screw, a transmission slider and a guide rod. A screw motor is arranged at one end of the transmission screw, and the setting direction of the transmission screw and the guide rod is consistent with the sliding direction of the transverse slide rail. The screw motor is connected and communicated with the controller; the transmission slider is matched with the transmission screw, and the transmission slider is connected to the garbage collection box. The rotation of the transmission screw is used as a driving force to slide along the guide rod. The screw transmission assembly can more smoothly drive the cleaning and dust collection assembly to move toward the two sides of the subway track to clean the garbage on both sides of the subway track.

[0010] Furthermore, the multiple vacuum cleaners are evenly distributed at the bottom of the garbage collection box, and at least one vacuum cleaner is arranged at the position of the garbage collection box located on the inner side of the subway track and the position of the garbage collection box located on the outer side of the subway track, so that the vacuum cleaner can handle the garbage around the subway track more comprehensively and the cleaning coverage is wider.

[0011] Furthermore, both ends of the garbage collection box are equipped with position lights to serve as a warning.

[0012] An intelligent subway track garbage identification and automatic cleaning method utilizes the above-mentioned device for garbage identification and cleaning, comprising the following steps: an image acquisition module acquires garbage image information and transmits it to an image processing module; the image processing module processes the received image to determine whether the acquired image information is garbage; when the acquired image information is confirmed to be garbage, the location coordinates of the garbage are calculated; the image processing module transmits the identified garbage location coordinates to a controller; after obtaining the garbage location coordinates, the controller calculates a cleaning path for an electric-controlled cleaning vehicle using a path planning algorithm of an autonomous route planning system.

[0013] The controller controls the drive components of the electric cleaning vehicle and the track operating platform through the planned cleaning path, and controls the electric cleaning vehicle to approach the garbage location coordinates while controlling the cleaning and dust collection components to move on the horizontal slide rail, so that the electric cleaning vehicle and the track operating platform cooperate with each other and the cleaning and dust collection components move to the garbage destination.

[0014] After arriving at the destination, the controller drives the cleaning and dust collection component to suck the identified garbage into the garbage collection box. When it detects that the garbage has been successfully cleaned, the controller turns off the cleaning and dust collection component and drives the electric cleaning vehicle to find the next piece of garbage.

[0015] Furthermore, the method for the image processing module to process the received image information includes the following steps:

[0016] The image processing module scales and binarizes the received image frames, gradually reduces the binarization threshold from 255, and uses the edge detection algorithm to extract the contour edge lines in the image.

[0017] Hough line detection is used to extract the straight lines in the binary threshold image and mask the rest.

[0018] The rest of the image is processed by adaptive binarization and Canny edge detection to obtain the edge closed area.

[0019] The area of ​​the closed region is screened, and the closed region with an area ranging from 10 to 1000 pixels is extracted, and feature extraction is performed to determine whether the image information collected by the image acquisition module is garbage.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] 1. The present invention uses a machine vision system to identify and locate garbage on the subway track, calculates the movement path based on the located position, and drives the electric-controlled cleaning vehicle through a controller to make it move forward along the subway track. At the same time, the controller controls the track operating platform to move the cleaning and dust collection component toward the side of the subway track. The mutual cooperation between the electric-controlled cleaning vehicle and the track operating platform can not only accurately clean the garbage at the located location, but also, by advancing and retreating the electric-controlled cleaning vehicle along the subway track and moving the track operating platform laterally along the subway track, the cleaning system can move over a large range within a two-dimensional plane, greatly expanding the cleaning range.

[0022] 2. The machine vision system of the present invention uses the OPENCV library to build an embedded system hardware platform, and adopts a recognition algorithm that combines image enhancement, adaptive binarization, Canny edge detection and convolutional neural network to realize the recognition and positioning of subway track garbage, with higher recognition accuracy and more accurate positioning coordinates.

[0023] 3. This invention can fully utilize the subway's nighttime maintenance period by rationally allocating the number of cleaning devices deployed, replacing manual labor in cleaning subway tracks. Furthermore, the deployed intelligent cleaning vehicles rely on the subway's "third rail" as their power supply platform, fully ensuring the stability and security of the power supply, as well as the design of a backup power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a schematic diagram of the structure of the cleaning device of the present invention placed on the subway track.

[0025] Figure 2 This is a flow chart of the system control of the present invention.

[0026] Among them, 1. Electric-controlled cleaning vehicle, 2. Image acquisition module, 3. Clearance lights, 4. Cleaning and dust collection components, 5. Garbage collection box, 6. Track operating platform, 8. Sliding wheel assembly, 9. Current collector, 10. Power supply system. DETAILED DESCRIPTION

[0027] The following combination Figure 1 and Figure 2 , a detailed description of the specific embodiments of the present invention is provided. In the description of the present invention, it should be understood that the terms "center," "upper," "lower," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like, indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or components referred to must have a specific direction, be constructed, or operate in a specific direction. Therefore, they should not be construed as limiting the present invention.

[0028] The terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, features specified as "first" or "second" may explicitly or implicitly include one or more of such features; and in the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0029] It should be noted that the circuit connections involved in the present invention all adopt conventional circuit connection methods and do not involve any innovation.

[0030] Example

[0031] like Figure 1As shown, an intelligent subway track cleaning device based on embedded machine vision includes: an electric-controlled cleaning vehicle 1, a machine vision system, a track operating platform 6, and a power supply system 10. The electric-controlled cleaning vehicle 1 is equipped with a drive component and can move forward and backward along the direction of the subway track; the machine vision system is arranged at the front of the electric-controlled cleaning vehicle 1, and the machine vision system serves as a carrier for the operation of the main program. The machine vision system includes: a controller, an image recognition system, and an autonomous route planning system. The controller is connected to the drive component of the electric-controlled cleaning vehicle 1. For example, the drive component of the electric-controlled cleaning vehicle 1 is a servo motor, and the controller is connected to the motor chip of the servo motor; the image recognition system The system comprises an image acquisition module 2 and an image processing module. The image acquisition module 2 is connected to the image processing module. In this embodiment, the image acquisition module 2 uses a high-definition camera for image acquisition, and the image processing module is connected to a controller for communication. An autonomous route planning system is connected to the controller for receiving image data processed by the image processing module and calculating the motion path of the electric-controlled cleaning vehicle 1 to the garbage location coordinates based on the processed image data. The autonomous route planning system achieves autonomous learning through interaction with the environment. Combined with a deep autoencoder learning algorithm, the system processes raw image data and autonomously extracts image features, improving the system's autonomy. Specifically, it includes path planning, obstacle detection and avoidance, navigation and positioning, decision-making and control, and real-time environmental perception. It is worth noting that the autonomous route planning system in this embodiment utilizes existing autonomous driving technology, such as Google's autonomous driving project, Waymo. Its autonomous navigation and planning system combines lidar, sensor fusion, real-time maps, and artificial intelligence algorithms for efficient autonomous planning and execution. Autoware is an open source autonomous driving software platform that includes complete path planning, positioning, and control modules. The track operating platform 6 includes: a transverse slide rail 7, a cleaning and dust collection component 4 and a garbage collection box 5, such as Figure 1 As shown, the transverse slide rail 7 is set above the electric-controlled cleaning vehicle 1 through a bracket, and the sliding direction is toward both sides of the subway track; the cleaning and dust collection component 4 is connected to the controller for communication, and the start and stop are controlled by the controller. The cleaning and dust collection component 4 includes a garbage collection box 5 and multiple vacuum cleaners. The garbage collection box 5 is slidably set on the transverse slide rail 7 through the sliding wheel assembly 8, and moves toward both sides of the subway track through the transverse slide rail 7. The sliding wheel assembly 8 includes a bracket and multiple rollers set on the bracket, and the bracket is connected to the garbage collection box 5; multiple vacuum cleaners are arranged at the bottom of the garbage collection box 5, and the vacuum cleaner's dust collection pipe is vertically set, and the output end of the dust collection pipe is connected to the garbage collection box 5; the power supply system 10 is used to provide power to the electric-controlled cleaning vehicle 1, the machine vision system and the track working platform 6.

[0032] In some embodiments, the power supply system 10 adopts a third rail power supply method. The third rail is an additional third rail for power supply in addition to the two rails on which the train runs. The collector of the electric train contacts and slides on the third rail, and transmits electricity to the train. Compared with overhead cables, it is cheaper and is widely used in subway systems. The power supply system 10 adopts a third rail power supply method. A current collector 9 is provided on the track operating platform 6. The current collector 9 is connected to the third rail. Compared with the existing cleaning device that directly installs an external power supply on the car body, power supply through the third rail can not only save the volume of the cleaning device and ensure the flexibility of the cleaning device, but also provide power quality assurance for the stable operation of the cleaning device, ensuring that the cleaning device can work in a relatively narrow subway track environment.

[0033] In some embodiments, the machine vision system utilizes the existing OPENCV library to build an embedded system hardware platform, employing a recognition algorithm that combines image enhancement, adaptive binarization, Canny edge detection, and convolutional neural networks to identify trash within subway tracks. This system can adapt to the complex environment of subway tracks, accurately and efficiently identifying small and light trash within the tracks, locating them, and performing path planning. This allows the track operating platform 6 and the electric-controlled cleaning vehicle 1 to move together, enabling unmanned and autonomous operation of the cleaning device.

[0034] In some embodiments, a screw drive assembly is configured on the transverse rail 7. The screw drive assembly comprises a drive screw, a drive slider, and a guide rod. The guide rod and the transverse rail 7 are arranged in the same orientation. A screw motor is configured at one end of the drive screw, which is in communication with a controller. The drive slider is coupled to the drive screw, which is connected to the waste collection box 5. The rotation of the drive screw provides a driving force, causing the waste collection box 5 and the multiple vacuum cleaners to slide along the guide rod. The screw drive assembly can more smoothly drive the cleaning and dust collection assembly 4 to move toward the sides of the subway track.

[0035] In some embodiments, multiple vacuum cleaners are evenly distributed at the bottom of the garbage collection box 5, and at least one vacuum cleaner is arranged at the position of the garbage collection box 5 located on the inner side of the subway track and the position of the garbage collection box 5 located on the outer side of the subway track. A larger cleaning area can be covered on both the inner and outer sides of the subway track through the movement of the transmission slider.

[0036] In some embodiments, both ends of the electric cleaning vehicle 1 and the garbage collection box 5 are equipped with position marker lights 3 for warning purposes.

[0037] An intelligent subway track garbage identification and automatic cleaning method, which uses the above-mentioned subway track cleaning device to identify and automatically clean garbage, includes the following steps:

[0038] The image acquisition module 2 collects garbage image information and transmits it to the image processing module. The image processing module processes the received image and determines whether the collected image information is garbage. When it is confirmed that the collected image information is garbage, the location coordinates of the garbage are calculated. The image processing module sends the location coordinates of the identified garbage to the controller; after the controller obtains the location coordinates of the garbage, it calculates the cleaning path of the electric-controlled cleaning vehicle 1 through the path planning algorithm of the autonomous route planning system.

[0039] The controller controls the driving components of the electric cleaning vehicle 1 and the track operating platform 6 through the planned cleaning path, and controls the electric cleaning vehicle 1 to approach the garbage location coordinates while controlling the cleaning and dust collection component 4 to move on the transverse slide rail 7, so that the electric cleaning vehicle 1 and the track operating platform 6 cooperate with each other, so that the cleaning and dust collection component 4 moves to the garbage destination.

[0040] After arriving at the destination, the controller drives the vacuum cleaner of the cleaning and dust collection component 4 to suck the identified garbage into the garbage collection box 5. When it is detected that the garbage has been successfully cleaned, the controller turns off the vacuum cleaner of the cleaning and dust collection component 4 and drives the electric cleaning vehicle 1 to find the next garbage.

[0041] The method for calculating the coordinates of the garbage location includes:

[0042] First, a self-calibration algorithm based on Zhang Zhengyou's calibration method is proposed to eliminate camera distortion. The algorithm combines adaptive binarization with Canny edge detection to extract the garbage outline. Then, based on monocular vision-based kinematic odometry, the algorithm converts the 3D world coordinate system to the camera coordinate system. Monocular vision-based kinematic odometry uses a single camera (monocular vision) to measure the distance between the object in front and the camera. The camera is calibrated using a fusion positioning algorithm based on GPS and the HMC5883L compass, overlaid with Baidu Maps API benchmarks to generate global geographic coordinates.

[0043] Based on Zhang's calibration principle, a self-calibration algorithm is proposed:

[0044] .

[0045] .

[0046] .

[0047] .

[0048] .

[0049] λ is an arbitrary scaling factor, r 1，r 2 It is the direction vector of the projection of the two coordinate axes of the image plane coordinate system in the world coordinate system, r 3 Represents the third column vector of the rotation matrix, h 1 is the first column vector of the homography matrix, h 2 is the second column vector of the homography matrix, h 3 is the third column vector of the homography matrix, M 1 represents the camera intrinsic parameter matrix, t is the translation vector, which represents the vector from the origin of the world coordinate system to the optical center of the camera.

[0050] The imaging principle of the camera:

[0051] .

[0052] d is the horizontal distance between the target and the camera, h is the height of the camera from the ground, y is the vertical coordinate of the target in the image, y 0 is the vertical coordinate of the center of the image, f is the focal length of the camera, α is the pitch angle of the camera.

[0053] The relationship between the projection coordinates (x, y) of the target point P on the image plane and the three-dimensional world coordinates (X, Y, Z) is:

[0054] .

[0055] .

[0056] Z is the depth component of the target point P in the camera coordinate system, that is, Z v .

[0057] .

[0058] Based on the kinematic ranging of the front target of monocular vision, the process of capturing the image by the camera is set as the process of projecting the three-dimensional world coordinates of the objective object to the coordinates stored in the image frame through coordinate transformation:

[0059] .

[0060] p It is the image frame coordinate of the object projected on the imaging plane, P is the coordinate of the object in the objective three-dimensional world, M1 is the camera intrinsic parameter matrix, which is only related to the internal parameters of the camera. M 2 is the external parameter matrix, which can be obtained through camera calibration. The image frame storage coordinates are obtained by the pixel coordinates of the target position in the captured image. The above formula can be used to calculate the coordinates of the object in the objective three-dimensional world. P .

[0061] The method for the image processing module to process the received image information includes the following steps:

[0062] The image processing module scales and binarizes the received image frames, gradually reduces the binarization threshold from 255, and uses the edge detection algorithm to extract the contour edge lines in the image.

[0063] Hough line detection is used to extract the straight lines in the binary threshold image and mask the rest.

[0064] The rest of the image is processed by adaptive binarization and Canny edge detection to obtain the edge closed area.

[0065] The closed areas are screened, with those between 10 and 1000 pixels being extracted for feature extraction. This determines whether the image information captured by the image acquisition module 2 is garbage. After training the machine vision system, the accuracy of garbage recognition in this embodiment reached 93.1%. This significantly improves the accuracy of garbage recognition in complex environments and, in conjunction with the track operating platform 6, enhances garbage removal efficiency.

[0066] The above disclosures are only several preferred specific embodiments of the present invention. However, the embodiments of the present invention are not limited thereto. Any changes that can be conceived by those skilled in the art should fall within the scope of protection of the present invention.

Claims

1. An intelligent subway track cleaning device based on embedded machine vision, characterized in that: include: An electric-controlled cleaning vehicle (1) is equipped with a drive assembly and is capable of moving forward and backward along the subway track; A track operating platform (6) comprising: a transverse slide rail (7) sliding toward both sides of the subway track, the transverse slide rail (7) being arranged above the electric-controlled cleaning vehicle (1); a cleaning and dust collection assembly (4) comprising: a garbage collection box (5) slidingly arranged on the transverse slide rail (7); a plurality of vacuum cleaners arranged at the bottom of the garbage collection box (5), with the vacuum cleaners having suction pipes arranged vertically, and the output ends of the suction pipes being connected to the garbage collection box (5); A machine vision system is provided on an electric-controlled cleaning vehicle (1), comprising: a controller connected to a drive assembly of the electric-controlled cleaning vehicle (1) and connected and communicated with a control module of a vacuum cleaner; an image recognition system comprising: an image acquisition module (2) and an image processing module, wherein the image acquisition module (2) is connected to the image processing module, and the image processing module is connected and communicated with the controller; an autonomous route planning system is connected and communicated with the controller, and is used to receive image data processed by the image processing module and calculate a motion path of the electric-controlled cleaning vehicle (1) to reach the coordinates of a garbage location based on the processed image data; The machine vision system uses the existing OPENCV library to build an embedded system hardware platform and adopts a recognition algorithm that combines image enhancement, adaptive binarization, Canny edge detection and convolutional neural network to identify garbage on the subway track. The transverse slide rail (7) is provided with a screw transmission assembly, the screw transmission assembly having a transmission screw, a transmission slider and a guide rod, one end of the transmission screw is provided with a screw motor, and the setting direction of the transmission screw and the guide rod is consistent with the sliding direction of the transverse slide rail (7), and the screw motor is connected to the controller for communication; the transmission slider is matched with the transmission screw, and the transmission slider is connected to the garbage collection box (5).

2. The intelligent subway track cleaning device based on embedded machine vision according to claim 1, characterized in that: Also includes: A power supply system (10) is used to provide power to the electric-controlled cleaning vehicle (1), the machine vision system, and the track operating platform (6). The power supply system (10) adopts a third-track power supply method. A current receiver (9) is provided on the track operating platform (6), and the current receiver (9) is connected to the third track.

3. The intelligent subway track cleaning device based on embedded machine vision according to claim 1, characterized in that: The plurality of vacuum cleaners are evenly arranged at the bottom of the garbage collection box (5), and at least one vacuum cleaner is arranged at both the position of the garbage collection box (5) located inside the subway track and the position of the garbage collection box (5) located outside the subway track.

4. The intelligent subway track cleaning device based on embedded machine vision according to claim 1, characterized in that: Both ends of the garbage collection box (5) are equipped with clearance lights (3).

5. An intelligent subway track garbage identification and automatic cleaning method, characterized in that: Using any one of claims 1 to 4 to identify and clean up garbage comprises the following steps: The image acquisition module (2) acquires garbage image information and transmits it to the image processing module. The image processing module processes the received image and determines whether the acquired image information is garbage. When the acquired image information is confirmed to be garbage, the location coordinates of the garbage are calculated. The image processing module sends the location coordinates of the identified garbage to the controller. After the controller obtains the location coordinates of the garbage, it calculates the cleaning path of the electric-controlled cleaning vehicle (1) through the autonomous route planning system. The controller controls the driving components of the electric-controlled cleaning vehicle (1) and the track operating platform (6) through the planned cleaning path, controls the electric-controlled cleaning vehicle (1) to approach the garbage location coordinates, and controls the cleaning and dust collection component (4) to move on the transverse slide rail (7), so that the electric-controlled cleaning vehicle (1) and the track operating platform (6) cooperate with each other, and the cleaning and dust collection component (4) moves to the garbage destination; After arriving at the destination, the controller drives the cleaning and dust collection component (4) to suck the identified garbage into the garbage collection box (5). When it is detected that the garbage has been successfully cleaned, the controller turns off the cleaning and dust collection component (4) and drives the electric-controlled cleaning vehicle (1) to search for the next garbage.

6. The intelligent subway track garbage identification and automatic cleaning method according to claim 5 is characterized in that: The method for the image processing module to process the received image information includes the following steps: The image processing module scales and binarizes the received image frames, gradually reduces the binarization threshold from 255, and uses the edge detection algorithm to extract the contour edge lines in the image; Extract the straight line in the binary threshold image and mask the rest; Adaptive binarization and Canny edge detection are performed on the remaining images to obtain edge closed areas; The area of ​​the closed region is screened, and the closed region with an area ranging from 10 to 1000 pixels is extracted, and feature extraction is performed to determine whether the image information collected by the image acquisition module (2) is garbage.

Images