A method for switching from laser navigation to QR code navigation

Abstract

The present invention relates to a method for switching from laser navigation to QR code navigation, belonging to the technical field of automatic guided vehicle or autonomous mobile robot navigation. The multi-line lidar periodically detects the ground reflectivity around the navigation switching point, calculates the change amount of the reflectivity, and preset a mutation threshold to determine whether the reflectivity mutation occurs, so as to evaluate whether the current environment is suitable for navigation switching. After confirming the environmental stability, the autonomous mobile vehicle continues to move towards the navigation switching point, and scans the ground QR code mark through the QR code sensor when approaching. After successful recognition, it switches from laser navigation to QR code navigation. After switching, a convolutional neural network model is used to predict the visual integrity of the QR code label, and whether to trigger a detour instruction or continue to drive along the predetermined path is determined according to the recognition result. The present invention effectively improves the stability and reliability of switching from laser navigation to QR code navigation, and is applicable to complex and changeable industrial environments.

Description

Technical Field

[0001] This application relates to the technical field of automatic guided vehicle or autonomous mobile robot navigation, and particularly relates to a method for switching from laser navigation to QR code navigation. Background Art

[0002] An automatic guided vehicle (AGV) is an intelligent wheeled mobile robot that uses electromagnetic or optical automatic guidance devices and travels along a specified guidance path to complete handling tasks. It has advantages such as fast movement, strong controllability, and good safety, and is widely used in factory automated production lines, warehousing logistics, and material transfer in airports and ports.

[0003] Laser SLAM (Simultaneous Localization and Mapping) navigation is a technology for robots to autonomously locate and build maps in unknown environments. The term "SLAM" itself reflects the characteristic of simultaneously performing localization and map building. Among them, "laser" refers to using a lidar (LiDAR) as the main sensor to obtain environmental information. The lidar determines the distance to surrounding objects by emitting laser beams and measuring the time of the reflected light, thereby generating point cloud data of the environment.

[0004] QR code navigation is achieved by laying QR codes on the ground along the AGV's travel path, and the AGV obtains its current position information by scanning these QR codes with a camera. The specific process is as follows: The camera on the AGV reads the QR code image laid on the ground and transmits the image coordinate information of the QR code to the AGV controller. The controller calculates these coordinate data, determines the position of the QR code on the map, and sends a navigation path instruction to the AGV trolley. The AGV trolley navigates according to the received path instruction and finally scans the destination QR code to confirm arrival when it reaches the destination.

[0005] Laser + QR code composite navigation is a composite navigation method that combines the advantages of laser navigation and QR code navigation. Laser navigation provides high-precision environmental perception and positioning capabilities, while QR codes provide clear position reference points for the robot. The combination of the two can further improve the positioning accuracy to meet millimeter-level or even higher accuracy requirements; at the same time, laser navigation can real-time sense changes in the surrounding environment and adjust the path planning in a timely manner. As an auxiliary positioning means, QR codes provide additional position calibration when laser navigation is interfered with or has errors, improving the stability and reliability of navigation.

[0006] However, when using laser + QR code composite navigation in the QR code navigation mode of laser cutting, the repeat positioning accuracy of laser navigation will deviate due to the influence of the external environment. At the same time, the scanning range of the QR code sensor is too small, resulting in the situation that the QR code sensor cannot scan the QR code at the navigation switching point during laser navigation, thus causing the system to stop and error, affecting the stability of the entire system.

[0007] Therefore, how to improve the stability and reliability of laser navigation switching to QR code navigation is an urgent problem to be solved. Summary of the Invention

[0008] The main purpose of this application is to provide a method for laser navigation to switch to QR code navigation, aiming to solve the problem of unstable navigation caused by sudden environmental changes or damaged QR code labels when laser navigation switches to QR code navigation in the prior art.

[0009] To achieve the above object, this application provides a method for laser navigation to switch to QR code navigation, including the following steps: S1. During the laser navigation stage, use a multi-line lidar to periodically detect the ground reflectivity within a radius of 5m around the navigation switching point. The navigation switching point is a key position point preset on the laser navigation path and connected to the QR code navigation area; S2. According to the ground reflectivity obtained in step S1, denoted as ; According to Calculate the change amount of the reflectivity of adjacent sampling points within a radius of 5m around the navigation switching point during the laser navigation stage ; Among them, Represents the reflectivity value of the I-th sampling point, and the change amount of the reflectivity Represents the absolute value of the reflectivity difference between each sampling point and its next point; S3. Preset a mutation threshold M, and compare the mutation threshold M with the change amount of the reflectivity Calculated in one detection period in step S2. When the mutation threshold M is less than the change amount of the reflectivity , it is considered that a reflectivity mutation has occurred between the sampling points And , and calculate the reflectivity mutation rate according to ; Among them, Represents the reflectivity mutation rate based on the mutation threshold M; Is the total number of reflectivity data points collected by the lidar in one detection period; Is a mutation indication function, and its definition is: ; When When , ; In other cases ; S4, according to the reflectivity mutation rate calculated in step S3, determine whether the current environment is suitable for navigation switching; if the reflectivity mutation rate If the reflectivity mutation rate is lower than the preset stability threshold, the environment is considered stable and suitable for navigation switching. If the value is higher than the stability threshold, the environment is considered unstable, the navigation switch is delayed and laser navigation is continued until the reflectivity mutation rate drops below the stability threshold; S5. After confirming that the environment is stable, the autonomous mobile vehicle continues to move towards the navigation switching point, and when the autonomous mobile vehicle approaches the navigation switching point, the QR code sensor starts to scan the QR code mark on the ground; if the QR code sensor successfully recognizes the QR code mark, the navigation mode is immediately switched from laser navigation to QR code navigation; S6. After switching to QR code navigation, the convolutional neural network model is used to predict the visual integrity of the QR code label. If the oil coverage area is >15% or physical damage is detected, the detour command is triggered 200ms in advance; if the oil coverage area is not >15% or physical damage is not detected, the autonomous mobile vehicle continues to travel along the predetermined path of the QR code navigation.

[0010] Preferably, the following steps are also included: the autonomous mobile vehicle performs QR code navigation switching through several QR codes regularly placed on the navigation switching point; the regular placement is specifically to arrange the main QR code label at the center position of the navigation switching point of the QR code navigation, and expand it according to a diamond topology with it as the geometric center.

[0011] Preferably, the step S6 specifically includes the following steps: S6-1, obtaining a QR code label, and denoising, graying, and contrast enhancing the QR code label, and locating the QR code area by contour extraction to obtain a preprocessed QR code image; S6-2, extracting the processed QR code image features layer by layer from the feature extraction part of the convolutional neural network model, and outputting a global feature vector; S6-3, inputting the global feature vector into the classification decision part of the convolutional neural network model, performing classification judgment, and outputting the recognition results of the oil coverage area and physical damage; S6-4, comparing the recognition result with the preset threshold, if the oil coverage area is greater than 15% or the physical damage exceeds the preset threshold, triggering a detour instruction and outputting a detour path suggestion; S6-5, if the threshold is not exceeded, the autonomous mobile vehicle continues to travel along the predetermined path of the QR code navigation.

[0012] Preferably, the steps for establishing the detour path specifically include the following steps: loading the detailed map information of the current area from the pre-stored environmental map, and determining the current position and attitude of the autonomous mobile vehicle in real time by fusing the data of the QR code sensor, IMU, and wheel encoder; converting the obtained map information into a weighted directed graph G(V, E), where V is the set of nodes and E is the set of edges, and taking the current vehicle position as the starting point C of the detour path, and selecting the next available QR code marker point or path node as the detour target point G; by the algorithm starts from the target node G, traces back to the starting node C through the parent node, records all the nodes on the path, and obtains the optimal path from the current position to the detour target point.

[0013] Preferably, the detailed map information is stored in the form of a grid map or a node map, where each node represents an accessible position, each edge represents the path between nodes, and the detailed map information includes at least the feasible area, obstacle positions, and alternative paths.

[0014] Preferably, in the weighted directed graph G(V, E), each node v ∈ V represents a position, each edge (u, v) ∈ E represents the path between two positions, and the weight w(u, v) of the edge represents the cost of the path, and the cost of the path is specifically any one of distance, time, and real-time traffic status.

[0015] Preferably, during the navigation process according to the QR code, the autonomous mobile vehicle continuously monitors the deviation between itself and the predetermined path, and uses the data provided by the inertial measurement unit for real-time correction to ensure that the autonomous mobile vehicle can accurately move along the predetermined path.

[0016] The beneficial effects of the technical solution of the present invention are as follows:

[0017] By real-time monitoring the change amount of the ground reflectivity, it is possible to accurately judge whether the current environment is suitable for switching the navigation mode, thereby avoiding switching during unstable environments, resulting in navigation failure or path deviation. And by using the multi-line lidar to periodically detect the density of dynamic obstacles and the ground reflectivity, combined with the calculation of the reflectivity mutation rate, it is possible to effectively identify the mutation situations in the environment.

[0018] After confirming the environmental stability, the autonomous mobile vehicle continues to move towards the navigation switching point, and starts to scan the ground QR code marker when approaching the navigation switching point, ensuring the navigation mode switching at the best timing, improving the accuracy and efficiency of navigation. And by using the convolutional neural network model to predict the visual integrity of the QR code label, potential problems such as excessive oil stain coverage or physical damage can be discovered and processed in advance, thereby triggering the detour instruction in time and avoiding possible navigation interruption or accidents.

[0019] By regularly placing a number of two-dimensional codes at the navigation switching points, the detection area of the two-dimensional codes is significantly expanded, enabling the two-dimensional code sensor of the autonomous mobile vehicle to more easily and quickly scan the two-dimensional code markers when approaching the navigation switching point, thereby reducing the probability of navigation switching delay or failure caused by two-dimensional code detection failure. Especially in complex environments (such as scenes with large light changes or uneven ground reflectivity), the arrangement of a number of two-dimensional codes can effectively improve the robustness of the system.

[0020] Meanwhile, after the autonomous mobile vehicle arrives at the navigation switching point under the guidance of the laser navigation system, the two-dimensional code navigation system accurately determines the starting point of the two-dimensional code navigation by scanning the two-dimensional code ID, ensuring the accuracy and smoothness of the navigation switching, and avoiding path deviation or vehicle stagnation problems caused by incorrect starting point judgment. At the same time, the system can re-plan the two-dimensional code navigation path according to the starting point to ensure that the vehicle can seamlessly connect and continue to move forward along the predetermined path after switching. Brief Description of the Drawings

[0021] Figure 1 is a flowchart in an embodiment of the present application;

[0022] Figure 2 is a two-dimensional code placement diagram in an embodiment of the present application.

[0023] The realization, functional features and advantages of the purpose of the present application will be further described in conjunction with the embodiments with reference to the drawings. Detailed Description of the Embodiments

[0024] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present application and should not be construed as a limitation of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0025] In addition, if the description in the present application involves "first", "second", etc., it is only for descriptive purposes (such as for distinguishing the same or similar elements), and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the fact that those of ordinary skill in the art can implement them. When the combination of the technical solutions appears to be contradictory or unable to be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present application.

[0026] See Figure 1 , the present invention proposes a method for switching from laser navigation to QR code navigation, including the following steps:

[0027] S1. In the laser navigation stage, the ground reflectivity within a radius of 5 m around the navigation switching point is periodically detected by a multi-line lidar, and the navigation switching point is a key position point preset on the laser navigation path and connected to the QR code navigation area;

[0028] S2. According to the ground reflectivity obtained in step S1, denoted as ; according to formula (1), the change amount of the reflectivity of adjacent sampling points within a radius of 5 m around the navigation switching point in the laser navigation stage is calculated as ;

[0029] Formula (1): .

[0030] Among them, represents the reflectivity value of the I-th sampling point. Therefore, the change amount of the reflectivity is specifically the absolute value of the difference in reflectivity between each sampling point and its next point.

[0031] S3. Preset a mutation threshold M, and compare the mutation threshold M with the change amount of the reflectivity calculated in step S2 within a detection period. When the mutation threshold M is less than the change amount of the reflectivity , it is considered that a reflectivity mutation has occurred between the sampling points and , and the reflectivity mutation rate is calculated according to formula (2);

[0032] Formula (2): .

[0033] Among them, represents the reflectivity mutation rate based on the mutation threshold M; is the total number of reflectivity data points collected by the lidar within a detection period; is a mutation indication function, and its definition is: ; when , ; in other cases ;

[0034] S4. According to the reflectivity mutation rate calculated in step S3, determine whether the current environment is suitable for navigation switching; specifically, if the reflectivity mutation rate is lower than the preset stability threshold, it is considered that the environment is stable and suitable for navigation switching; if the reflectivity mutation rate If it is higher than the stability threshold, the environment is considered unstable, navigation switching is delayed, and laser navigation is continued until the reflectivity mutation rate drops below the stability threshold;

[0035] S5. After confirming that the environment is stable, the autonomous mobile vehicle continues to move toward the navigation switching point, and when the autonomous mobile vehicle approaches the navigation switching point, the QR code sensor starts to scan the QR code mark on the ground. If the QR code sensor successfully recognizes the QR code mark, the navigation mode is immediately switched from laser navigation to QR code navigation;

[0036] S6. After switching to QR code navigation, a convolutional neural network model is used to predict the visual integrity of the QR code label. If the oil coverage area is greater than 15% or physical damage is detected, a detour command is triggered 200ms in advance. If the oil coverage area is not greater than 15% or physical damage is not detected, the autonomous mobile vehicle continues to travel along the predetermined path of the QR code navigation.

[0037] In this embodiment, by real-time monitoring of the change in ground reflectivity, it is possible to accurately determine whether the current environment is suitable for switching navigation modes, thereby avoiding switching when the environment is unstable, resulting in navigation failure or path deviation. Multi-line laser radar is used to periodically detect dynamic obstacle density and ground reflectivity, combined with the calculation of reflectivity mutation rate, to effectively identify mutations in the environment.

[0038] Furthermore, after confirming that the environment is stable, the autonomous mobile vehicle continues to move toward the navigation switching point and begins to scan the ground QR code mark when approaching the navigation switching point, ensuring that the navigation mode is switched at the best time, improving the accuracy and efficiency of navigation. The convolutional neural network model predicts the visual integrity of the QR code label, and can detect and deal with potential problems in advance, such as excessive oil coverage or physical damage, thereby triggering detour instructions in time to avoid possible navigation interruptions or accidents.

[0039] In one embodiment, the following steps are also included:

[0040] The autonomous mobile vehicle performs QR code navigation switching through a number of QR codes regularly placed at the navigation switching point; the regular placement specifically arranges the main QR code label at the center position of the navigation switching point of the QR code navigation, and expands it in a diamond topology with it as the geometric center.

[0041] In this embodiment, by regularly placing a number of two-dimensional codes at the navigation switching points, the detection area of the two-dimensional codes is significantly expanded, enabling the two-dimensional code sensor to more easily and quickly scan the two-dimensional code markers when the autonomous mobile vehicle approaches the navigation switching point, thereby reducing the probability of navigation switching delay or failure caused by two-dimensional code detection failure. Especially in complex environments (such as scenes with large light changes or uneven ground reflectivity), the arrangement of a number of two-dimensional codes can effectively improve the robustness of the system.

[0042] Meanwhile, after the autonomous mobile vehicle arrives at the navigation switching point under the guidance of the laser navigation system, the two-dimensional code navigation system accurately determines the starting point of the two-dimensional code navigation by scanning the two-dimensional code ID, ensuring the accuracy and smoothness of the navigation switching, and avoiding path deviation or vehicle stagnation problems caused by incorrect starting point judgment. At the same time, the system can re-plan the two-dimensional code navigation path according to the starting point to ensure that the vehicle can seamlessly connect and continue to move forward along the predetermined path after switching.

[0043] Specifically, referring to Figure 2 , where point 374 is the navigation switching point and also the target point of the laser navigation; normally, the autonomous mobile vehicle arrives at point 374 under the guidance of the laser navigation system, and the two-dimensional code navigation system scans the two-dimensional code at point 374 after the laser navigation system completes the navigation and continues to guide the vehicle.

[0044] When the laser navigation system deviates due to external environmental influences and fails to guide the vehicle to the two-dimensional code at point 374 and shifts to any direction in front of, behind, to the left, or to the right of point 374, it will scan one of the two-dimensional code points placed in the four directions of point 373.

[0045] Since the autonomous mobile vehicle always scans one of the two-dimensional code points around 374, the two-dimensional code navigation system avoids the risk of downtime caused by code dropout; the two-dimensional code navigation system re-determines the position of the autonomous mobile vehicle and the starting point of the two-dimensional code navigation according to the scanned two-dimensional code ID, and plans a new two-dimensional code navigation path according to the starting point and guides the vehicle to continue moving forward.

[0046] In one of the embodiments, step S6 specifically includes the following steps: S6-1, obtain a QR code label, perform denoising, grayscale conversion, and contrast enhancement on the QR code label, and locate the QR code area by contour extraction to obtain a preprocessed QR code image; S6-2, layer by layer extract the features of the processed QR code image from the feature extraction part of the convolutional neural network model, and output a global feature vector; S6-3, input the global feature vector into the classification decision part of the convolutional neural network model for classification judgment, and output the recognition results of the oil stain coverage area and physical damage; S6-4, compare the recognition results with a preset threshold. If the oil stain coverage area is greater than 15% or the physical damage exceeds the preset threshold, trigger a detour instruction and output a detour path suggestion; S6-5, if not exceeding the threshold, the autonomous mobile vehicle continues to travel along the predetermined path of QR code navigation.

[0047] In this embodiment, by performing denoising, grayscale conversion, and contrast enhancement on the collected image, the clarity and quality of the image are significantly improved, thereby ensuring that subsequent feature extraction is more accurate and reliable, effectively reducing the influence of environmental light changes, camera noise, etc. on the image quality, and providing a stable basis for the entire navigation switching process. At the same time, using the convolutional neural network model to automatically extract the feature vector of the image eliminates the need for manual intervention, improves the processing efficiency and consistency, quickly and accurately captures the key information in the QR code image, and provides strong support for subsequent classification judgment.

[0048] On the other hand, inputting the extracted feature vector into the classification decision part of the fully convolutional neural network model for classification judgment can accurately output the recognition results of the oil stain coverage area and physical damage, enabling it to timely detect and handle potential navigation risks and ensure the safe operation of the autonomous mobile vehicle. Further, when the recognition result shows that the oil stain coverage area is greater than 15% or the physical damage exceeds the preset threshold, the system not only triggers a detour instruction but also outputs a detour path suggestion, effectively avoiding navigation failure or path deviation caused by environmental problems, and at the same time providing clear detour guidance for the autonomous mobile vehicle to ensure that it can quickly restore the correct navigation state. When the recognition result shows that the environmental conditions meet the requirements, the autonomous mobile vehicle will continue to travel along the predetermined path of QR code navigation to ensure the continuity and stability of the navigation system in most cases.

[0049] In one of the embodiments, the steps for establishing the detour path specifically include the following steps:

[0050] The system loads the detailed map information of the current area from a pre-stored environmental map, and determines the current position and attitude of the autonomous mobile vehicle in real time by fusing the data of the QR code sensor, IMU (Inertial Measurement Unit) and wheel encoder; wherein, the detailed map information at least includes a feasible area, obstacle positions, and alternative paths; and the detailed map information is stored in the form of a grid map or a node map, where each node represents an accessible position and each edge represents a path between nodes.

[0051] The obtained map information is converted into a weighted directed graph G(V, E), where V is the set of nodes and E is the set of edges, and the current vehicle position is used as the starting point C of the detour path, and the next available QR code marker point or path node is selected as the detour target point G; in the weighted directed graph G(V, E), each node v ∈ V represents a position, each edge (u, v) ∈ E represents a path between two positions, and the weight w(u, v) of the edge represents the cost of the path, and the cost of the path is specifically any one of distance, time, and real-time traffic status.

[0052] Through The algorithm starts from the target node G, traces back to the starting node C through the parent nodes, records all the nodes on the path, and obtains the optimal path from the current position to the detour target point.

[0053] In this embodiment, by converting the map information into a weighted directed graph and combining The algorithm for path planning significantly improves the navigation ability and adaptability of the autonomous mobile vehicle in a complex environment. This technology is efficient, accurate, and robust, can dynamically respond to environmental changes, reduce computational complexity, and support multi-objective optimization, providing strong technical support for the high-precision navigation requirements in industrial automation.

[0054] In one of the embodiments, during the navigation according to the QR code, the autonomous mobile vehicle continuously monitors the deviation between it and the predetermined path, and uses the data provided by the inertial measurement unit (IMU) for real-time correction to ensure that the vehicle can accurately move along the predetermined path.

[0055] It should be noted that in this text, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article or method for laser navigation to switch to QR code navigation including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such a process, apparatus, article or method for laser navigation to switch to QR code navigation. Without further limitation, an element defined by the statement "including one..." does not exclude the presence of additional identical elements in the process, apparatus, article or method for laser navigation to switch to QR code navigation including such element.

[0056] The above are only the preferred embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. A method for switching from laser navigation to QR code navigation, characterized in that, Applied to an autonomous mobile vehicle, the method comprises the following steps: S1, in the laser navigation stage, periodically detecting the ground reflectivity within a radius of 5m around a navigation switching point by means of a multi-line laser radar, wherein the navigation switching point is a key position point pre-set on the laser navigation path and connected to the QR code navigation area; S2. According to the ground reflectivity obtained in step S1, it is recorded as ;according to The calculated reflectivity change of adjacent sampling points within a radius of 5m around the navigation switching point during the laser navigation phase is ; in, It is represented as the reflectivity value of the Ith sampling point, and the reflectivity change It is represented by the absolute value of the reflectivity difference between each sampling point and the next point; S3, preset the mutation threshold M, and compare the mutation threshold M with the reflectivity change within a detection cycle calculated in step S2 In contrast, when the mutation threshold M is less than the reflectivity change , then it is considered that at the sampling point and There was a sudden change in reflectivity between Calculate the reflectivity mutation rate ;in, represents the reflectivity mutation rate based on the mutation threshold M; is the total number of reflectivity data points collected by the LiDAR in one detection cycle; is the mutation indicator function, which is defined as: ;when hour, ; In other cases ; S4, according to the reflectivity mutation rate calculated in step S3, determine whether the current environment is suitable for navigation switching; if the reflectivity mutation rate If the reflectivity mutation rate is lower than the preset stability threshold, the environment is considered stable and suitable for navigation switching. If the value is higher than the stability threshold, the environment is considered unstable, the navigation switch is delayed and laser navigation is continued until the reflectivity mutation rate drops below the stability threshold; S5. After confirming that the environment is stable, the autonomous mobile vehicle continues to move towards the navigation switching point, and when the autonomous mobile vehicle approaches the navigation switching point, the QR code sensor starts to scan the QR code mark on the ground; if the QR code sensor successfully recognizes the QR code mark, the navigation mode is immediately switched from laser navigation to QR code navigation; S6. After switching to QR code navigation, the convolutional neural network model is used to predict the visual integrity of the QR code label. If the oil coverage area is >15% or physical damage is detected, the detour command is triggered 200ms in advance; if the oil coverage area is not >15% or physical damage is not detected, the autonomous mobile vehicle continues to travel along the predetermined path of the QR code navigation.

2. A method for switching from laser navigation to QR code navigation according to claim 1, characterized in that, The following steps are also included: The autonomous mobile vehicle performs QR code navigation switching through a number of QR codes regularly placed at the navigation switching point; the regular placement specifically arranges the main QR code label at the center position of the navigation switching point of the QR code navigation, and expands it in a diamond topology with it as the geometric center.

3. A method for switching from laser navigation to QR code navigation according to claim 1, characterized in that, The step S6 specifically includes the following steps: S6-1, obtaining a QR code label, and denoising, graying, and contrast enhancing the QR code label, and locating the QR code area by contour extraction to obtain a preprocessed QR code image; S6-2, extracting the processed QR code image features layer by layer from the feature extraction part of the convolutional neural network model, and outputting a global feature vector; S6-3, inputting the global feature vector into the classification decision part of the convolutional neural network model, performing classification judgment, and outputting the recognition results of the oil coverage area and physical damage; S6-4, comparing the recognition result with the preset threshold, if the oil coverage area is greater than 15% or the physical damage exceeds the preset threshold, triggering a detour instruction and outputting a detour path suggestion; S6-5, if the threshold is not exceeded, the autonomous mobile vehicle continues to travel along the predetermined path of the QR code navigation.

4. A method for switching from laser navigation to two-dimensional code navigation according to claim 3, characterized in that, The steps for establishing the detour path specifically include the following steps: Load the detailed map information of the current area from the pre-stored environmental map, and determine the current position and orientation of the autonomous mobile vehicle in real time by fusing the data of the QR code sensor, IMU, and wheel encoder; Convert the obtained map information into a weighted directed graph G(V, E), where V is the set of nodes and E is the set of edges, and use the current vehicle position as the starting point C of the detour path, and select the next available QR code marker point or path node as the detour target point G; Through The algorithm starts from the target node G, traces back to the starting node C through the parent node, records all the nodes on the path, and obtains the optimal path from the current position to the detour target point.

5. A method for switching from laser navigation to two-dimensional code navigation according to claim 4, characterized in that, The detailed map information is stored in the form of a grid graph or a node graph, wherein each node represents a reachable location, each edge represents a path between nodes, and the detailed map information at least includes a feasible area, obstacle locations, and alternate paths.

6. The method for switching from laser navigation to two-dimensional code navigation according to claim 5, wherein In the weighted directed graph G(V,E), each node v∈V represents a location, each edge (u,v)∈E represents a path between two locations, and the edge weight w(u,v) represents the cost of the path, which is specifically any one of distance, time and real-time traffic status.

7. A method for switching from laser navigation to QR code navigation according to claim 1, characterized in that, While driving according to the QR code navigation, the deviation between the autonomous mobile vehicle and the predetermined path is continuously monitored, and real-time corrections are made using data provided by the inertial measurement unit to ensure that the autonomous mobile vehicle can move accurately along the predetermined path.

Images