Road construction road roller intelligent identification and path planning system based on machine vision

Abstract

The invention discloses a road construction road roller intelligent identification and path planning system based on machine vision, and the system comprises the following components: an image collection module which comprises a high-definition camera which is installed at the front side of a road roller and covers a construction area in front of and around the road roller, the target recognition module is used for recognizing a road roller and an obstacle in a road construction scene by applying a deep learning-based target recognition algorithm, and the path planning module is used for performing intelligent path planning according to the position information of the road roller and the obstacle provided by the target recognition module, generating an optimal road roller driving path in real time, and transmitting the optimal road roller driving path to the road construction scene. And the data processing and control module adopts a high-performance processor. Through image acquisition and target identification, construction safety is guaranteed, intelligent path planning and dynamic adjustment improve the construction efficiency, the data processing and communication module ensures system stability and remote management, data storage and analysis optimize performance, and the intelligent level and quality of road construction are integrally improved.

Description

Technical Field

[0001] The present invention relates to the technical field of roller identification and path planning systems, and in particular to a machine vision-based intelligent identification and path planning system for road construction rollers. Background Art

[0002] With the acceleration of urbanization and the continuous advancement of infrastructure construction, road construction projects are increasing, and the requirements for road construction quality and efficiency are becoming higher and higher. In road construction, the road roller is one of the key construction equipment, and its operation quality and efficiency directly affect key indicators such as road flatness and compaction.

[0003] The traditional road construction roller operation method mainly relies on the operator's experience and on-site judgment. However, this method has many limitations in the complex construction site environment. First, the construction site environment is complex and changeable, and there are various obstacles, such as construction workers, other construction equipment, stones, etc. It is difficult for operators to fully grasp the surrounding situation in real time, and collision accidents are prone to occur, which not only endangers the safety of personnel and equipment, but may also cause construction interruptions and affect the construction progress.

[0004] Secondly, there is a lack of accurate identification and monitoring methods for different types of rollers and their posture and position information during construction. This makes it difficult for the construction management department to accurately evaluate and dispatch the operation of the rollers, unable to achieve refined management of the construction process, and difficult to ensure the consistency and stability of construction quality.

[0005] In terms of path planning, the traditional method is usually to make a rough plan based on the construction drawings and the experience of the operators, which lacks real-time and flexibility. When unexpected situations occur at the construction site, such as new obstacles or changes in construction conditions, the path cannot be adjusted in time, resulting in low construction efficiency and even the need for repeated operations, increasing construction costs and time costs.

[0006] In addition, with the development of science and technology, remote monitoring and management are being used more and more widely in various fields. However, in the field of road construction rollers, the existing systems are insufficient in terms of stability and security of data transmission, as well as convenience of remote control and parameter adjustment of the system, and cannot meet the needs of modern construction management.

[0007] In summary, the traditional road construction roller operation method can no longer meet the requirements of today's efficient, safe and precise construction. There is an urgent need for an intelligent recognition and path planning system based on machine vision to improve the intelligence level of roller operation, ensure construction safety, improve construction efficiency and quality, and realize refined management and remote monitoring of the construction site. Summary of the invention

[0008] The present invention proposes a machine vision-based intelligent recognition and path planning system for road construction rollers to solve the problems mentioned in the above-mentioned prior art.

[0009] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: a machine vision-based intelligent identification and path planning system for road construction rollers, comprising the following components: Image acquisition module: Contains a high-definition camera with a resolution of at least 1080P and a frame rate of 30fps or above. It can adapt to the lighting conditions of the road construction site, including direct strong light and low light environments, and has an automatic dimming function.

[0010] The camera is installed on the front side of the roller, covering the construction area in front of and around the roller, with a viewing angle of not less than 120 degrees. It is connected to the roller through a stable mounting bracket and shock-absorbing device to reduce the impact of roller vibration on image acquisition.

[0011] It has image preprocessing functions, including denoising and contrast enhancement. It uses image filtering algorithms to remove noise in images and enhances image contrast through histogram equalization technology, making the collected images clearer and more accurate, which is conducive to subsequent image processing and analysis.

[0012] Target recognition module: Uses deep learning-based target recognition algorithms to identify road rollers and obstacles in road construction scenes, including construction workers, other construction equipment, and stones.

[0013] For the identification of road rollers, it can accurately identify their model, posture and location information. By learning and analyzing the appearance characteristics of the road rollers, it can accurately distinguish different models of road rollers. At the same time, it uses image processing technology to calculate the posture angle of the road roller and the specific location coordinates at the construction site.

[0014] For obstacle recognition, it has multi-category recognition capabilities, classifies and identifies obstacles based on their shape, size, and color characteristics, and determines their location and size information, providing accurate data support for subsequent path planning.

[0015] The target recognition module contains a deep neural network model with multiple convolutional layers and fully connected layers. It is trained by annotated data to optimize the model parameters and improve the accuracy and speed of target recognition. The target recognition accuracy is no less than 95%.

[0016] Introduce the target recognition confidence formula: ,in represents the confidence of target recognition, is the number of detected targets, It is The classification probability of each target (output by the model), is an indicator function. If The classification of the target is correct. ,otherwise This formula can be used to evaluate the reliability of target recognition, and when the confidence level is below a certain threshold, a re-detection or manual confirmation mechanism can be triggered to improve the accuracy of recognition.

[0017] Path planning module: Based on the location information of the roller and obstacles provided by the target recognition module, combined with the map information of the construction site and construction requirements, intelligent path planning is performed to generate the optimal roller driving path in real time, taking into account factors including the roller's minimum turning radius, construction efficiency, and safety, to ensure that the roller can efficiently complete the road construction task while avoiding obstacles.

[0018] The path planning module also has the function of dynamic path adjustment. When new obstacles or changes in the situation appear at the construction site, the path can be replanned in time. The response time of path adjustment does not exceed 1 second to ensure the continuous and stable operation of the roller. Communicate with the control system of the roller, transmit the planned path information to the roller, and guide the roller to drive according to the planned path. The communication method adopts reliable wireless communication technology, such as Bluetooth and Wi-Fi, to ensure the stability and real-time performance of data transmission. Introduce the path planning safety assessment formula: ,in represents the safety evaluation index of the path, , , is the weight coefficient, is the minimum distance between the roller and the nearest obstacle, is the relative speed between the roller and the obstacle (calculated considering the motion state of the roller and the obstacle), is the maximum possible acceleration of the roller on the path (estimated based on roller performance and path conditions). This formula is used to comprehensively evaluate the safety of the planned path to ensure that the path selection is within the safety range. When the safety evaluation index is lower than the safety threshold, the path is replanned to improve construction safety.

[0019] Data processing and control module: It uses a high-performance processor with a main frequency of no less than 2.0GHz and multi-core processing capabilities. It is responsible for processing and coordinating the data of the image acquisition module, target recognition module and path planning module.

[0020] The running memory is not less than 4GB and the storage capacity is not less than 32GB, which is used to store system programs, intermediate data and historical records. It uses fast storage media, such as SSD solid state drives, to improve data reading and writing speeds and system operating efficiency.

[0021] Real-time monitoring and control of the operation of the entire system, including monitoring the working status of each module, fault diagnosis and processing. When a module failure or abnormal situation occurs, appropriate measures can be taken in time, such as restarting the module and switching to a backup module, to ensure the stability and reliability of the system.

[0022] It has a human-computer interaction interface, which displays the system's operating status, image acquisition results, target recognition information and path planning solutions on the display screen. Operators can set parameters, start and stop the system through the interface. The interface design is simple, intuitive and easy to operate.

[0023] Power management module: provides stable power supply for the entire system, supports a wide voltage input range, such as 12V-24V, and is suitable for the on-board power system of the road roller.

[0024] It has power supply stabilization and filtering functions, which can effectively remove noise and fluctuations in the power supply, ensure that all modules of the system work in a stable power supply environment, and prevent system failure or performance degradation caused by power supply problems.

[0025] It has a low-power management mode. When the system is idle or does not need to run at full power, it can automatically reduce the power consumption of some modules and extend the battery life of the system. For example, when the roller stops working, it automatically reduces the frame rate of the image acquisition module and the operating frequency of the processor to reduce power consumption.

[0026] Communication module: includes a module for communicating with the remote monitoring center, supports 4G / 5G network communication, and can upload real-time data of the construction site, such as video images, roller operating status, and path planning information, to the remote monitoring center.

[0027] It has data encryption function and uses advanced encryption algorithm to encrypt the transmitted data to ensure the security and confidentiality of the data and prevent the data from being stolen or tampered with during transmission.

[0028] The remote monitoring center can remotely control and adjust the parameters of the system. The remote monitoring personnel can send instructions through the network to modify the operating parameters of the system, such as adjusting the target recognition threshold and changing the path planning strategy. At the same time, they can obtain feedback information from the system in real time to achieve remote monitoring and management of the construction site.

[0029] Furthermore, the system has efficient image acquisition performance. In a strong direct light environment, the image brightness uniformity error does not exceed 10%. In a weak light environment, the image signal-to-noise ratio is not less than 30dB, to ensure that the quality of the acquired image can meet the subsequent processing requirements under different lighting conditions.

[0030] Furthermore, the target recognition module has a certain stability difference in the recognition accuracy of different types of obstacles in a complex construction site environment. The recognition accuracy for common construction personnel and large construction equipment is not less than 98%, and the recognition accuracy for small stone irregular obstacles is not less than 90%, so as to adapt to the diversity and complexity of obstacles in actual construction scenarios.

[0031] Furthermore, the path planning module introduces the construction efficiency evaluation formula when considering the construction efficiency factor: ,in Indicates construction efficiency, It is the shortest time theoretically required to complete the construction task (calculated based on the performance of the roller and the amount of construction tasks). It is the time spent in the actual construction process. This formula is used to evaluate the effect of path planning on improving construction efficiency. It is required that under normal construction conditions, the construction efficiency should be improved by no less than 15%.

[0032] Furthermore, the system has good anti-interference capabilities. In the presence of electromagnetic interference at the construction site, the data transmission bit error rate does not exceed 1‰, ensuring stable and reliable data transmission between system modules and with the remote monitoring center.

[0033] Furthermore, a method for applying a road construction roller intelligent identification and path planning system based on machine vision comprises the following steps: S1. Image acquisition: The high-definition camera of the image acquisition module collects image data of the road construction site in real time, and pre-processes the collected images, including denoising and contrast enhancement operations, to improve image quality.

[0034] S2. Target recognition: The target recognition module uses a deep learning algorithm to analyze the preprocessed images, identify the roller and various obstacles, and determine their location, posture, and category information.

[0035] S3. Path planning: The path planning module uses intelligent algorithms to generate the optimal driving path for the roller based on the information provided by the target recognition module, combined with the map of the construction site and construction requirements, and dynamically adjusts the path in real time according to changes at the construction site.

[0036] S4. Data transmission and control: The data processing and control module transmits the path planning information to the control system of the roller, and uploads the system's operating status and related data to the remote monitoring center through the communication module, thereby realizing real-time control and remote monitoring of the roller.

[0037] S5. System monitoring and maintenance: Real-time monitoring of the system's operating status, fault diagnosis and processing of each module, regular software upgrades and hardware maintenance of the system to ensure system stability and reliability.

[0038] Furthermore, a method for intelligent identification and path planning of a road construction roller further includes the following steps: S6. Data storage and analysis: Store the image data, target recognition results, and path planning information during the system operation process, establish a historical database, and optimize the target recognition algorithm and path planning strategy through analysis of historical data to improve the performance and adaptability of the system.

[0039] S7. Emergency handling: When the system detects an abnormal situation, such as the sudden appearance of a large number of obstacles or a roller failure, the emergency handling mechanism is activated, the operation of the roller is immediately stopped or switched to manual control mode, and an alarm message is sent to the remote monitoring center. At the same time, detailed information of the abnormal situation is recorded for subsequent analysis and processing.

[0040] S8. User authority management: Set different levels of user authority, including administrator authority and ordinary user authority. Administrators can set, modify system parameters and view all data, while ordinary users can only perform basic operations and view part of the data, ensuring the security of the system and the confidentiality of data.

[0041] Furthermore, in the path planning process, the path smoothness evaluation formula is introduced: ,in represents the path smoothness, is the number of points on the path, It is the first The coordinates of the points are used to evaluate the smoothness of the planned path through this formula. The path smoothness is required to be within a certain range to ensure the smoothness of the roller's travel and reduce the wear on the roller's mechanical parts and the impact on the construction quality.

[0042] Compared with the prior art, the present invention has the following beneficial effects: The present invention uses advanced image acquisition modules and target recognition modules to achieve accurate recognition of road rollers and obstacles on the construction site. It can accurately identify the model, posture and position of the road roller, as well as the type and location information of obstacles such as construction personnel, equipment, and stones, with high target recognition accuracy. This greatly improves the safety of the construction site, effectively reduces the risk of collision accidents, ensures the safety of personnel and equipment, and provides an accurate data basis for subsequent path planning, making the operation of the road roller more orderly and controllable.

[0043] The path planning module combines multiple factors for intelligent planning and has dynamic adjustment functions. By considering the minimum turning radius of the roller, construction efficiency and safety, the optimal driving path is generated in real time, and it can be quickly replanned when the situation changes, and the path adjustment response time is short. This significantly improves construction efficiency, reduces construction time and costs, and ensures that the roller can efficiently complete road construction tasks, avoiding construction interruptions and repeated operations caused by unreasonable paths or emergencies.

[0044] The data processing and control modules, communication modules and related performance guarantee measures in the system ensure the stable operation of the entire system and the reliability of data transmission. It can monitor the system status in real time, diagnose and handle faults, realize remote monitoring and management, and facilitate the construction management department to control and dispatch the construction site in real time. It can adjust system parameters and strategies in time, improving the refinement and convenience of construction management.

[0045] In addition, the data storage and analysis function in the system further improves the performance and adaptability of the system by optimizing algorithms and strategies through the analysis of historical data, enabling the system to be continuously optimized and improved with use, better adapting to different construction site environments and needs, and providing strong support for the intelligent and high-quality development of road construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is an overall schematic block diagram of the intelligent identification and path planning system for road construction rollers based on machine vision proposed by the present invention; Figure 2 This is a schematic block diagram of a path planning module of the machine vision-based intelligent identification and path planning system for road construction rollers proposed by the present invention. DETAILED DESCRIPTION

[0047] 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.

[0048] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise" and "counterclockwise" indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0049] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, and it can be the internal connection of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0050] Reference Figure 1-Figure 2 :A road construction roller intelligent identification and path planning system based on machine vision, including the following components: Image acquisition module: Contains a high-definition camera with a resolution of at least 1080P and a frame rate of 30fps or above. It can adapt to the lighting conditions of the road construction site, including direct strong light and low light environments, and has an automatic dimming function.

[0051] The camera is installed on the front side of the roller, covering the construction area in front of and around the roller, with a viewing angle of not less than 120 degrees. It is connected to the roller through a stable mounting bracket and shock-absorbing device to reduce the impact of roller vibration on image acquisition.

[0052] It has image preprocessing functions, including denoising and contrast enhancement. It uses image filtering algorithms to remove noise in images and enhances image contrast through histogram equalization technology, making the collected images clearer and more accurate, which is conducive to subsequent image processing and analysis.

[0053] Target recognition module: Uses deep learning-based target recognition algorithms to identify road rollers and obstacles in road construction scenes, including construction workers, other construction equipment, and stones.

[0054] For the identification of road rollers, it can accurately identify their model, posture and location information. By learning and analyzing the appearance characteristics of the road rollers, it can accurately distinguish different models of road rollers. At the same time, it uses image processing technology to calculate the posture angle of the road roller and the specific location coordinates at the construction site.

[0055] For obstacle recognition, it has multi-category recognition capabilities, classifies and identifies obstacles based on their shape, size, and color characteristics, and determines their location and size information, providing accurate data support for subsequent path planning.

[0056] The target recognition module uses a deep learning-based target recognition algorithm to identify rollers and obstacles in road construction scenes, including construction workers, other construction equipment, stones, etc.; by learning and analyzing the appearance features of the rollers, it can accurately identify rollers of different models, and use image processing technology to calculate the roller's posture angle and specific location coordinates at the construction site.

[0057] For obstacle recognition, it has multi-category recognition capabilities, classifies and identifies obstacles based on their shape, size, and color characteristics, and determines their location and size information.

[0058] Contains a deep neural network model that uses an architecture based on MobileNetV3 and is optimized: A multi-scale feature fusion module is introduced to fuse feature maps of different layers. The idea of ​​feature pyramid network (FPN) is adopted to fuse feature maps with rich high-level semantic information with feature maps with high spatial resolution of low layers, and construct a feature representation with multi-scale information to better identify objects of different sizes and shapes.

[0059] Combining the channel attention module (CAM) and the spatial attention module (SAM), the CAM automatically focuses on the channel features that contribute more to target recognition by learning the importance weight of each channel; the SAM focuses on the key spatial positions in the image, suppresses information in irrelevant areas, and highlights the key areas related to the roller and obstacles in the image, thereby improving the recognition accuracy and efficiency of the model.

[0060] The trained model is compressed and accelerated, and the knowledge of the complex teacher model is transferred to the student model using knowledge distillation technology. The model is quantized and the floating-point parameters are converted into integers to reduce the storage and calculation amount of the model and improve the running speed.

[0061] The target recognition module contains a deep neural network model with multiple convolutional layers and fully connected layers. It is trained by annotated data to optimize the model parameters and improve the accuracy and speed of target recognition. The target recognition accuracy is no less than 95%.

[0062] Introduce the target recognition confidence formula: ,in represents the confidence of target recognition, is the number of detected targets, It is The classification probability of each target (output by the model), is an indicator function. If The classification of the target is correct. ,otherwise This formula can be used to evaluate the reliability of target recognition, and when the confidence level is below a certain threshold, a re-detection or manual confirmation mechanism can be triggered to improve the accuracy of recognition.

[0063] The path planning module is the core component of the intelligent identification and path planning system for road construction rollers based on machine vision. Its main function is to generate the optimal roller driving path in real time based on a variety of input information to ensure the efficiency and safety of construction. The following is a detailed description of the technical solution of this module: The path planning module is closely connected to the target recognition module and receives the location information of the roller and obstacles provided by it in real time. The target recognition module uses image processing and deep learning algorithms to accurately identify the coordinate position of the roller at the construction site and the specific positions of various obstacles (such as construction personnel, other construction equipment, stones, etc.), and transmits this information to the path planning module in a standardized data format.

[0064] The location information of the roller includes not only its plane coordinates (x, y), but also its attitude information (such as heading angle, tilt angle, etc.), which is crucial for accurately planning the driving path of the roller. The path planning module will update and store this information in real time so that it can be called at any time during the path planning process.

[0065] The path planning module performs intelligent path planning based on the location information of the roller and obstacles provided by the target recognition module, combined with the map information and construction requirements of the construction site, and generates the optimal driving path for the roller in real time: A global path planning method based on the fusion of the A algorithm and the Dijkstra algorithm is adopted. In the map preprocessing stage, the Dijkstra algorithm is used to calculate and store the shortest distance from each grid point to the target point. The A algorithm refers to these pre-calculated distance information when searching to improve the search speed and efficiency. At the same time, the heuristic function of the A* algorithm is optimized to make it more accurately estimate the actual distance to the target point, reduce the number of search nodes, and improve the accuracy of path planning. The path is optimized considering the construction requirements, such as giving priority to planning paths that pass through areas that require a higher degree of compaction, avoiding planning paths with frequent turns and sudden braking, and dynamically adjusting the path planning strategy according to the construction stage and task priority.

[0066] Relying on high-precision construction site maps, the system uses a variety of technical means such as lidar scanning, satellite image positioning, and on-site measurement to build and update the map in real time. The map contains the accurate location and shape information of obstacles and road terrain information; the map is layered, with different layers representing different types of information and semantically labeled to better analyze and process map information for path planning.

[0067] With the function of dynamic path adjustment, when new obstacles or changes in the situation appear at the construction site, the path can be replanned in time, and the response time of path adjustment does not exceed 1 second: A real-time monitoring system is built through multi-sensor fusion (such as millimeter-wave radar, camera, ultrasonic sensor, etc.) to perceive changes in the construction site in real time. Based on the event-driven programming model, the current path planning is immediately suspended and the re-planning process is started when changes are detected. Using reinforcement learning-based algorithms (such as deep Q network DQN and its improved version), the road roller, as an intelligent agent, learns the optimal path planning strategy by interacting with the environment. Taking time factors into consideration, it predicts changes in the construction site in the future and plans the path accordingly.

[0068] At the same time, the module receives map information of the construction site, which can be a pre-imported digital map or a real-time map generated through on-site measurement and modeling. The map contains detailed information such as the geometry of the road, slope, and obstacle distribution. The path planning module will merge the location information of the roller and obstacles with the map information and map it into a unified coordinate system, providing a basis for subsequent path planning calculations.

[0069] Constrain and optimize the path planning in combination with construction requirements. Construction requirements include but are not limited to compaction sequence, compaction times, construction progress requirements, etc. For example, for some road construction projects, it may be required to compact the edge of the road first and then gradually move towards the middle; or it may be required to complete the construction task of a certain area within a specific time period.

[0070] The path planning module will analyze and quantify these construction requirements and convert them into specific path planning constraints and optimization goals. For example, the compaction sequence requirements are converted into path sequence constraints, and the construction progress requirements are converted into constraints on path length and construction time. In this way, it is ensured that the generated path meets the actual needs of the construction.

[0071] A graph search-based algorithm is used to find the optimal path. First, the map of the construction site is abstracted into a graph structure, where nodes represent key locations in the map (such as road intersections, curve centers, boundary points of the construction area, etc.), and edges represent feasible paths between nodes. The path planning module constructs this graph structure based on map information and constraints such as the minimum turning radius of the roller.

[0072] In the graph search process, heuristic search algorithms such as A algorithm or Dijkstra algorithm are used. A algorithm selects the next node to be searched by comprehensively considering the actual cost from the starting node to the current node and the estimated cost from the current node to the target node. It can improve the search efficiency while ensuring that the optimal path is found. Dijkstra algorithm is a classic shortest path algorithm. It starts from the starting node and gradually expands outward until the target node is found. It is suitable for solving the shortest path problem in undirected graphs or weighted graphs.

[0073] During the path search, the algorithm constrains the path based on the minimum turning radius of the roller. When calculating a path from one node to another, it checks whether there is a path with a turning radius smaller than the minimum turning radius of the roller. If so, the path is excluded. This ensures that the generated path is a path that the roller can actually travel.

[0074] To improve construction efficiency, the path planning module considers multiple factors when generating a path. First, the theoretical time required to complete the entire construction task is estimated based on the total amount of construction tasks and the working efficiency of the roller. Then, during the path planning process, the driving distance and number of turns of the roller are minimized to reduce driving time and energy consumption.

[0075] For example, when selecting a route, it is preferred to choose a shorter and smoother route to avoid unnecessary detours and repeated driving. At the same time, considering the difference in driving speed of the roller under different road conditions, the path planning module will optimize the path according to the road information in the map (such as flatness, slope, etc.). On sections with good flatness and small slopes, the driving speed of the roller can be appropriately increased to shorten the construction time; on sections with poor road conditions, the speed should be appropriately reduced to ensure construction quality and safety.

[0076] In addition, the module will arrange the working order of the rollers according to the actual situation of the construction site. For example, if there are multiple construction areas that need to be compacted, the area that is closer to the current location and less difficult to construct will be selected for construction first, and then gradually move to other areas to reduce the roller's idle driving time and transfer time.

[0077] Safety is one of the important considerations for path planning. The path planning module evaluates the safety of the generated path in real time, mainly considering the distance between the roller and obstacles, relative speed, and other safety factors at the construction site.

[0078] For each potential path point, the module calculates the minimum distance between the roller and the surrounding obstacles. If it is found that the obstacles near a path point are too close and may threaten the driving safety of the roller, the module will automatically adjust the path to avoid these dangerous areas. For example, by adding a penalty term for obstacle distance during the graph search process, the algorithm tends to choose a path that is farther away from obstacles.

[0079] At the same time, considering the motion state of the roller and the obstacle, the path planning module calculates the relative speed between them in real time. If it is predicted that the relative speed between the roller and the obstacle is large on a certain path segment, which may increase the risk of collision, the module will take corresponding measures, such as slowing down or adjusting the path. For example, a safety buffer zone can be set in the path planning. When the roller enters the safety buffer zone, the driving speed is automatically reduced, or the path is replanned to avoid potential dangers.

[0080] In addition, the module will also consider other safety factors at the construction site, such as the traffic conditions at the construction site, the range of activities of construction workers, etc. If it is found that a certain path may conflict with the traffic flow at the construction site, or pass through an area with dense construction workers, the module will adjust the path in time to ensure the safety of the roller.

[0081] After generating the initial path, the path planning module will continue to monitor changes in the construction site. Through real-time communication with the image acquisition module and the object recognition module, it obtains information about the roller position, changes in obstacle positions, and other changes in construction conditions.

[0082] For example, if a new obstacle suddenly appears on the planned path or an existing obstacle moves while the roller is driving, the image acquisition module will capture these changes in time and transmit the information to the path planning module. At the same time, the position and posture information of the roller will also be updated in real time so that the path planning module can accurately grasp the current status of the roller.

[0083] When changes occur at the construction site and the original planned path is no longer applicable, the path planning module will immediately start the path replanning mechanism and recalculate the optimal path based on the latest roller position, obstacle location information and construction requirements.

[0084] During the path replanning process, the previous path planning results and existing map information will be fully utilized to improve the efficiency of replanning. For example, the affected path segments can be partially adjusted based on the original path instead of completely recalculating the entire path. At the same time, the current driving state and speed of the roller will be taken into account to ensure that the replanned path can be smoothly transitioned and will not cause excessive impact on the driving of the roller.

[0085] The response time of path adjustment is one of the important indicators to measure the performance of the path planning module. In order to ensure that the roller can respond to path changes in time, the path planning module will use efficient algorithms and data structures to complete path replanning in the shortest time. Generally, the response time of path adjustment does not exceed 1 second to ensure the continuous and stable operation of the roller.

[0086] The optimal path generated by the path planning module needs to be transmitted to the roller control system in real time to guide the roller to drive along the planned path. Reliable wireless communication technologies such as Bluetooth and Wi-Fi are used between the module and the roller control system to ensure the stability and real-time performance of data transmission.

[0087] When transmitting the path information, the path will be sent to the roller control system in the form of a series of coordinate points and driving instructions. The roller control system will control the roller's steering, driving speed and other parameters based on this information, so that the roller can accurately drive along the planned path. At the same time, the path planning module will maintain real-time communication with the roller control system and receive feedback from the roller at any time, such as the actual driving position and driving status, so as to adjust the path planning strategy in a timely manner.

[0088] The path planning module also has the function of dynamic path adjustment. When new obstacles or changes in the situation appear at the construction site, the path can be replanned in time. The response time of path adjustment does not exceed 1 second to ensure the continuous and stable operation of the roller. Communicate with the control system of the roller, transmit the planned path information to the roller, and guide the roller to drive according to the planned path. The communication method adopts reliable wireless communication technology, such as Bluetooth and Wi-Fi, to ensure the stability and real-time performance of data transmission. Introduce the path planning safety assessment formula: ,in represents the safety evaluation index of the path, , , is the weight coefficient, is the minimum distance between the roller and the nearest obstacle, is the relative speed between the roller and the obstacle (calculated considering the motion state of the roller and the obstacle), is the maximum possible acceleration of the roller on the path (estimated based on roller performance and path conditions). This formula is used to comprehensively evaluate the safety of the planned path to ensure that the path selection is within the safety range. When the safety evaluation index is lower than the safety threshold, the path is replanned to improve construction safety.

[0089] Data processing and control module: It uses a high-performance processor with a main frequency of no less than 2.0GHz and multi-core processing capabilities. It is responsible for processing and coordinating the data of the image acquisition module, target recognition module and path planning module.

[0090] The running memory is not less than 4GB and the storage capacity is not less than 32GB, which is used to store system programs, intermediate data and historical records. It uses fast storage media, such as SSD solid state drives, to improve data reading and writing speeds and system operating efficiency.

[0091] Real-time monitoring and control of the operation of the entire system, including monitoring the working status of each module, fault diagnosis and processing. When a module failure or abnormal situation occurs, appropriate measures can be taken in time, such as restarting the module and switching to a backup module, to ensure the stability and reliability of the system.

[0092] It has a human-computer interaction interface, which displays the system's operating status, image acquisition results, target recognition information and path planning solutions on the display screen. Operators can set parameters, start and stop the system through the interface. The interface design is simple, intuitive and easy to operate.

[0093] Power management module: provides stable power supply for the entire system, supports a wide voltage input range, such as 12V-24V, and is suitable for the on-board power system of the road roller.

[0094] It has power supply stabilization and filtering functions, which can effectively remove noise and fluctuations in the power supply, ensure that all modules of the system work in a stable power supply environment, and prevent system failure or performance degradation caused by power supply problems.

[0095] It has a low-power management mode. When the system is idle or does not need to run at full power, it can automatically reduce the power consumption of some modules and extend the battery life of the system. For example, when the roller stops working, it automatically reduces the frame rate of the image acquisition module and the operating frequency of the processor to reduce power consumption.

[0096] The communication module plays a vital role in the intelligent identification and path planning system of road construction rollers based on machine vision. It is the key link to achieve data interaction and remote management between the system and the remote monitoring center. The following is a detailed description of the technical solution of the communication module: The communication module integrates 4G and 5G network communication functions to adapt to the network environment and requirements of different construction sites. For 4G communication, it supports mainstream LTE standards, including FDD-LTE and TDD-LTE frequency bands, and can work stably under the widely covered 4G network. For 5G communication, it supports high-speed data transmission in the sub-6GHz frequency band, with higher bandwidth and lower latency characteristics to meet application scenarios with high real-time requirements.

[0097] The module is equipped with a high-performance 4G / 5G modem, which can achieve fast network connection and data transmission. During the network connection process, the module will automatically search for available network signals and select the best network connection based on factors such as signal strength, network quality and operator priority. At the same time, it has a network switching function. When the current network signal is unstable or interrupted, it can automatically switch to other available networks to ensure the continuity of data transmission.

[0098] In order to optimize data transmission efficiency, the communication module uses advanced wireless communication protocol stacks and data transmission optimization algorithms. For example, based on the TCP / IP protocol stack, the data transmission window size, congestion control algorithm, etc. are optimized to improve the data transmission speed and reliability in wireless networks. At the same time, it supports MIMO (multiple input multiple output) technology, which sends and receives data simultaneously through multiple antennas, further improving the transmission performance and stability of the network.

[0099] In terms of data transmission protocols, a combination of HTTP (Hypertext Transfer Protocol) and MQTT (Message Queue Telemetry Transmission) is adopted. For video images and roller operating status data with high real-time requirements, HTTP protocol is used for real-time push transmission to ensure that the remote monitoring center can obtain the latest data in a timely manner. For some non-real-time but important information, such as path planning information and system configuration parameters, MQTT protocol is used for publish / subscribe mode transmission to reduce network bandwidth occupancy and improve data transmission reliability.

[0100] In terms of data format, JSON (JavaScript Object Notation) is used as the main data exchange format. JSON format has the advantages of being concise, easy to read, and easy to parse, which facilitates data transmission and processing between different systems and platforms. The data collected at the construction site, such as frame data of video images, operating parameters of the roller (including position, speed, posture, pressure, etc.), coordinate points and strategy information of path planning, etc., are structured and encapsulated into JSON format data packets for transmission. At the same time, when receiving instructions and parameter adjustment information from the remote monitoring center, the data in JSON format can also be parsed and applied to the system.

[0101] The communication module uses advanced encryption algorithms to ensure data security and confidentiality. For data encryption, the AES (Advanced Encryption Standard) algorithm is selected, which has high-intensity encryption performance and good compatibility and is widely used in various security fields. During the transmission process, the data is encrypted in real time, the plaintext data is grouped according to a certain block size, and then the AES algorithm is used to encrypt each data block to generate ciphertext data for transmission.

[0102] In order to further enhance the security of data, a key management mechanism is also used in the encryption process. The system will regularly generate new encryption keys and distribute them to the communication module and remote monitoring center in a secure way. The transmission and storage of keys are encrypted to ensure the security of the keys. At the same time, a key rotation strategy is adopted to regularly replace encryption keys to reduce the risk of key cracking.

[0103] In addition to encrypting data, the communication link is also encrypted and protected. The SSL / TLS (Secure Sockets Layer / Transport Layer Security) protocol is used to establish a secure communication link. Before data transmission, the communication module and the remote monitoring center will conduct an SSL / TLS handshake process to negotiate encryption algorithms, keys and other parameters, and then establish an encrypted communication channel to ensure that data will not be stolen or tampered with during transmission.

[0104] In order to ensure the integrity of the transmitted data at the receiving end, the communication module uses message authentication code (MAC) technology. When sending data, a MAC value is generated based on the data content and the key, and it is sent together with the data. After receiving the data, the receiving end will recalculate the MAC value and compare it with the received MAC value. If the two MAC values ​​are consistent, it means that the data has not been tampered with during the transmission process. Otherwise, it is considered that the data has been damaged or tampered with, and the data will be requested to be resent.

[0105] At the same time, data verification mechanisms such as CRC (cyclic redundancy check) are also used in the data transmission process. By performing CRC calculation on the data, a check code is generated and added to the data frame for transmission. After receiving the data, the receiving end will perform CRC check on the data. If the check result is incorrect, it means that an error has occurred in the data transmission process, and corresponding error correction measures will be taken, such as requesting to resend the data or using error correction codes to repair the data.

[0106] The remote monitoring center sends instructions to the communication module through the network to achieve remote control and parameter adjustment of the system. The transmission of instructions uses the same network communication protocol and encryption method as data transmission to ensure the safe transmission of instructions. The format of the instruction adopts a custom instruction set format, including information such as instruction type, parameter list, and checksum. For example, for the instruction to adjust the target recognition threshold, it can be defined as "TARGET_RECOGNITION_THRESHOLD_ADJUST, [threshold_value], [checksum]", where "TARGET_RECOGNITION_THRESHOLD_ADJUST" is the instruction type, indicating the adjustment of the target recognition threshold; "[threshold_value]" is the specific threshold parameter; "[checksum]" is the checksum, which is used to verify the integrity of the instruction.

[0107] After receiving the instruction, the communication module will parse and verify the instruction. First, verify whether the checksum of the instruction is correct to ensure that the instruction has not been tampered with during transmission. Then, the specific parameters are parsed according to the instruction type and passed to the corresponding module for processing. For example, for the instruction to adjust the target recognition threshold, the communication module will pass the threshold parameter to the target recognition module, and the target recognition module will readjust the parameters of its recognition algorithm according to the new threshold parameter to optimize the target recognition accuracy.

[0108] In order to achieve remote monitoring and management, the communication module is also responsible for transmitting the system's feedback information to the remote monitoring center in real time. The system will regularly collect various feedback information, including the real-time operating status of the roller (such as current position, driving speed, compaction degree, etc.), the working status of the image acquisition module (such as whether the camera is working properly, image quality, etc.), the results of target recognition (such as the type and number of obstacles identified, etc.) and the execution of path planning (such as whether it is driving according to the planned path, whether it encounters obstacles, etc.).

[0109] The communication module organizes and encapsulates the feedback information and sends it to the remote monitoring center in accordance with the specified data format and transmission protocol. After receiving the feedback information, the remote monitoring center can understand the situation on the construction site in real time and conduct further analysis and processing as needed. For example, if it is found that the compaction degree of the roller does not meet the requirements, the remote monitoring personnel can adjust the working parameters of the roller by sending instructions, or notify the on-site operator to check and handle it.

[0110] When implementing remote control and parameter adjustment functions, the communication module also integrates permission management and security verification mechanisms in consideration of the security and reliability of the system. Users of the remote monitoring center need to undergo identity authentication and permission verification before they can send instructions to operate the system. The communication module interacts with the user management system of the remote monitoring center to obtain the user's identity information and permission level, and verify the user's operation request.

[0111] Only users with corresponding permissions can adjust specific system parameters or perform specific operations. For example, administrator users have the highest permissions and can set and modify all system parameters, while ordinary users may only be able to view the system's operating status and some basic information, and cannot adjust key parameters. Through this permission management mechanism, malicious operations on the system by illegal users are effectively prevented, ensuring the security and stability of the system.

[0112] In the present invention, the system has efficient image acquisition performance. In a strong direct light environment, the image brightness uniformity error does not exceed 10%, and in a weak light environment, the image signal-to-noise ratio is not less than 30dB, to ensure that the quality of the collected image can meet the subsequent processing requirements under different lighting conditions.

[0113] In the present invention, the target recognition module has a certain stability difference in the recognition accuracy of different types of obstacles in a complex construction site environment. The recognition accuracy for common construction personnel and large-scale construction equipment is not less than 98%, and the recognition accuracy for small stone irregular obstacles is not less than 90%, so as to adapt to the diversity and complexity of obstacles in actual construction scenes.

[0114] In the present invention, the path planning module introduces a construction efficiency evaluation formula when considering the construction efficiency factor: ,in Indicates construction efficiency, It is the shortest time theoretically required to complete the construction task (calculated based on the performance of the roller and the amount of construction tasks). It is the time spent in the actual construction process. This formula is used to evaluate the effect of path planning on improving construction efficiency. It is required that under normal construction conditions, the construction efficiency should be improved by no less than 15%.

[0115] In the present invention, the system has good anti-interference ability. In the presence of electromagnetic interference at the construction site, the data transmission bit error rate does not exceed 1‰, ensuring that the data transmission between the modules of the system and between the system and the remote monitoring center is stable and reliable.

[0116] The present invention also discloses a method for applying a road construction roller intelligent identification and path planning system based on machine vision, comprising the following steps: S1. Image acquisition: The high-definition camera of the image acquisition module collects image data of the road construction site in real time, and pre-processes the collected images, including denoising and contrast enhancement operations, to improve image quality.

[0117] S2. Target recognition: The target recognition module uses a deep learning algorithm to analyze the preprocessed images, identify the roller and various obstacles, and determine their location, posture, and category information.

[0118] S3. Path planning: The path planning module uses intelligent algorithms to generate the optimal driving path for the roller based on the information provided by the target recognition module, combined with the map of the construction site and construction requirements, and dynamically adjusts the path in real time according to changes at the construction site.

[0119] S4. Data transmission and control: The data processing and control module transmits the path planning information to the control system of the roller, and uploads the system's operating status and related data to the remote monitoring center through the communication module, thereby realizing real-time control and remote monitoring of the roller.

[0120] S5. System monitoring and maintenance: Real-time monitoring of the system's operating status, fault diagnosis and processing of each module, regular software upgrades and hardware maintenance of the system to ensure system stability and reliability.

[0121] The present invention also discloses a method for intelligent identification and path planning of a road construction roller, which further includes the following steps: S6. Data storage and analysis: Store the image data, target recognition results, and path planning information during the system operation process, establish a historical database, and optimize the target recognition algorithm and path planning strategy through analysis of historical data to improve the performance and adaptability of the system.

[0122] S7. Emergency handling: When the system detects an abnormal situation, such as the sudden appearance of a large number of obstacles or a roller failure, the emergency handling mechanism is activated, the operation of the roller is immediately stopped or switched to manual control mode, and an alarm message is sent to the remote monitoring center. At the same time, detailed information of the abnormal situation is recorded for subsequent analysis and processing.

[0123] S8. User authority management: Set different levels of user authority, including administrator authority and ordinary user authority. Administrators can set, modify system parameters and view all data, while ordinary users can only perform basic operations and view part of the data, ensuring the security of the system and the confidentiality of data.

[0124] In the process of path planning, the present invention introduces a path smoothness evaluation formula: ,in represents the path smoothness, is the number of points on the path, It is the first The coordinates of the points are used to evaluate the smoothness of the planned path through this formula. The path smoothness is required to be within a certain range to ensure the smoothness of the roller's travel and reduce the wear on the roller's mechanical parts and the impact on the construction quality.

[0125] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can replace or change the technical solution and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A road construction roller intelligent identification and path planning system based on machine vision, characterized in that: It includes the following components: Image acquisition module: It includes a high-definition camera with a resolution of at least 1080P and a frame rate of 30fps or above to adapt to different light intensity environmental conditions at the road construction site; the camera is installed on the front side of the roller, covering the construction area in front of and around the roller, with a viewing angle of not less than 120 degrees. The image acquisition module has image preprocessing functions, including denoising and contrast enhancement. It uses image filtering algorithms to remove noise in the image and enhances the image contrast through histogram equalization technology; Target recognition module: Uses deep learning-based target recognition algorithms to identify rollers and obstacles in road construction scenes, including construction personnel, other construction equipment, and stones. By learning and analyzing the appearance features of rollers, it can accurately identify rollers of different models, and uses image processing technology to calculate the roller's posture angle and specific location coordinates at the construction site. For obstacle recognition, it has multi-category recognition capabilities, classifies and identifies obstacles based on their shape, size, and color characteristics, and determines their location and size information; The target recognition module includes a deep neural network model with multiple convolutional layers and fully connected layers. The model is trained by annotated data to optimize the parameters of the model and improve the accuracy and speed of target recognition. The target recognition accuracy is not less than 95%. The multi-layer convolutional layer is a convolutional neural network architecture based on MobileNetV3. The convolution operation is decomposed into two steps: deep convolution and point-by-point convolution. Improvements are made on the basis of the residual structure. Channel expansion is performed first, followed by deep convolution, and finally channel compression. The h-swish activation function is used to improve the nonlinear expression ability of the network. Based on the convolutional neural network architecture of MobileNetV3, the Feature Pyramid Network (FPN) is integrated to fuse the high-level semantic information feature map with the low-level feature map with high spatial resolution, thereby constructing a feature representation with multi-scale information. Introduce the target recognition confidence formula: ,in represents the confidence of target recognition, is the number of detected targets, It is The classification probability of a target, is an indicator function. If The classification of the target is correct. ,otherwise ,When the confidence level is lower than the threshold, a re-test or manual confirmation mechanism can be triggered; Path planning module: Based on the location information of the roller and obstacles provided by the target recognition module, combined with the map information and construction requirements of the construction site, intelligent path planning is performed to generate the optimal driving path for the roller in real time; The path planning module also has the function of dynamic path adjustment. When new obstacles or changes in the situation appear at the construction site, the path can be re-planned in time. The response time of path adjustment does not exceed 1 second. It communicates with the control system of the roller, transmits the planned path information to the roller, and guides the roller to drive according to the planned path. The communication method adopts wireless communication technology, including Bluetooth and Wi-Fi. The path planning module performs path planning based on the location information of the roller and obstacles provided by the target recognition module, combined with the map information and construction requirements of the construction site, and generates the optimal driving path of the roller in real time, specifically: A global path planning method based on the fusion of the A algorithm and the Dijkstra algorithm is adopted. In the map preprocessing stage, the Dijkstra algorithm is used to calculate and store the shortest distance from each grid point to the target point. The A algorithm refers to these pre-calculated distance information when searching, and at the same time optimizes the heuristic function of the A* algorithm to further estimate the actual distance to the target point; priority is given to planning paths that pass through areas that require a higher degree of compaction, and paths with frequent turns and sudden braking are reduced. The path planning strategy is dynamically adjusted according to the construction stage and task priority; Based on the construction site map, laser radar scanning, satellite image positioning and on-site measurement are used to build and update the map in real time. The map contains the accurate location and shape information of obstacles and road terrain information. The map is layered, with different layers representing different types of information and semantically annotated. When new obstacles or changes in the situation appear at the construction site, the path is replanned, and the response time for path adjustment does not exceed 1 second: A real-time monitoring system is built through the fusion of millimeter-wave radar, camera, and ultrasonic multi-sensor to perceive changes on the construction site in real time. Based on the event-driven programming model, the current path planning is immediately suspended and the re-planning process is started when changes are detected. The deep Q network DQN algorithm based on reinforcement learning is adopted. The road roller, as an intelligent agent, learns the optimal path planning strategy through interaction with the environment. Introduce the path planning safety assessment formula: ,in represents the safety evaluation index of the path, , , is the weight coefficient, is the minimum distance between the roller and the nearest obstacle, is the relative speed between the roller and the obstacle, is the maximum acceleration of the roller on the path; Data processing and control module: It uses a high-performance processor with a main frequency of no less than 2.0GHz and multi-core processing capabilities, and is responsible for processing and coordinating the data of the image acquisition module, target recognition module, and path planning module; The running memory is not less than 4GB, and the storage capacity is not less than 32GB, which is used to store system programs, intermediate data and historical records, and uses SSD solid state hard disk storage media; Real-time monitoring and control of the operation of the entire system, including monitoring the working status of each module, fault diagnosis and processing, and taking measures to restart the module or switch to a backup module when a module failure or abnormal situation occurs; Power management module: provides stable power supply for the entire system, supports a wide voltage input range of 12V-24V, has power supply voltage stabilization and filtering functions, and removes noise and fluctuations in the power supply; Communication module: includes a module for communicating with the remote monitoring center, supports 4G / 5G network communication, and can upload real-time data of the construction site, including video images, roller operating status, and path planning information, to the remote monitoring center.

2. The machine vision-based intelligent identification and path planning system for road construction rollers according to claim 1 is characterized in that: The system has efficient image acquisition performance. In a direct light environment, the image brightness uniformity error does not exceed 10%. In a non-direct light environment, the image signal-to-noise ratio is not less than 30dB.

3. The machine vision-based intelligent identification and path planning system for road construction rollers according to claim 1 is characterized in that: In a complex construction site environment, the target recognition module has different stability in its recognition accuracy for different types of obstacles. The recognition accuracy for construction personnel and large construction equipment is no less than 98%, and the recognition accuracy for small stone and irregular obstacles is no less than 90%.

4. The machine vision-based intelligent identification and path planning system for road construction rollers according to claim 1 is characterized in that: The path planning module introduces the construction efficiency evaluation formula when considering the construction efficiency factor: ,in Indicates construction efficiency, It is the shortest time required to complete the construction task in theory. It is the time spent in the actual construction process.

5. The machine vision-based intelligent identification and path planning system for road construction rollers according to claim 1, characterized in that: In the presence of electromagnetic interference at the construction site, the data transmission bit error rate does not exceed 1‰, ensuring stable and reliable data transmission between system modules and with the remote monitoring center.

6. A method for intelligent identification and path planning of a road construction roller based on machine vision, used in the intelligent identification and path planning system for a road construction roller based on machine vision as claimed in any one of claims 1 to 5, characterized in that: The following steps are involved: S1. Image acquisition: The image data of the road construction site is collected in real time through the high-definition camera of the image acquisition module, and the collected images are pre-processed, including denoising and contrast enhancement operations, to improve image quality; S2. Target recognition: The target recognition module uses deep learning algorithms to analyze the preprocessed images, identify the road roller and various obstacles, and determine their location, posture, and category information; S3. Path planning: The path planning module uses intelligent algorithms to generate the optimal driving path for the roller based on the information provided by the target recognition module, combined with the map of the construction site and construction requirements, and dynamically adjusts the path in real time according to changes in the construction site; S4, data transmission and control: the data processing and control module transmits the path planning information to the control system of the roller, and uploads the system's operating status and related data to the remote monitoring center through the communication module, so as to realize real-time control and remote monitoring of the roller; S5. System monitoring and maintenance: Real-time monitoring of the system's operating status, fault diagnosis and processing of each module, software upgrades and hardware maintenance of the system to ensure system stability and reliability.

7. The method for intelligent identification and path planning of a road construction roller according to claim 6, characterized in that: The following steps are also included: S6. Data storage and analysis: Store the image data, target recognition results, and path planning information during the system operation process, establish a historical database, and optimize the target recognition algorithm and path planning strategy through the analysis of historical data; S7, emergency handling: When the system detects an abnormal situation, including a large number of obstacles or roller failure, the emergency handling mechanism is activated, the operation of the roller is immediately stopped or switched to manual control mode, and an alarm message is sent to the remote monitoring center, while recording detailed information of the abnormal situation; S8. User authority management: Set different levels of user authority, including administrator authority and ordinary user authority. Administrators can set, modify and view all data of the system parameters, while ordinary users can only perform basic operations and view some data.

8. The method for intelligent identification and path planning of a road construction roller according to claim 6, characterized in that: In the path planning process, the path smoothness evaluation formula is introduced: ,in represents the path smoothness, is the number of points on the path, It is the first The coordinates of a point.

Images