An intelligent fertilization vehicle path planning method suitable for a greenhouse environment

Abstract

The application relates to the field of intelligent path planning, and discloses an intelligent fertilizing vehicle path planning method suitable for a greenhouse environment, a road surface obstacle risk value evaluation mechanism based on an expert system and AHP (analytic hierarchy process) level analysis, road surface point cloud data established by combining binocular vision algorithm and deep learning technology, and comprehensive evaluation of road surface obstacle risks is realized. The obstacle avoidance mechanism of the traditional path planning method is effectively improved, so that the fertilizing vehicle can effectively deal with the path planning scene of a muddy path and a plant-shielded path in the greenhouse. The application improves the accessibility, timeliness and safety of the path planning of the greenhouse fertilizing vehicle.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent path planning, in particular to a path planning method for an intelligent fertilizing vehicle suitable for a greenhouse environment. BACKGROUND

[0002] It is of great urgency and importance to carry out the design of fertilizer blending and application equipment and improve the efficiency of chemical fertilizer blending.

[0003] At present, the greenhouse environment in China is complex and changeable, and various plants are planted. The ordinary intelligent operation machine faces the problems of difficult travel, heavy program burden of path planning algorithm, long time-consuming, low efficiency and high operation cost. For example, the measures taken by the ordinary automatic car to face the road sections with pits, protrusions or branches are to bypass the obstacles, and the measures taken to face the muddy road sections are to directly drive through. This will cause a lot of time to be wasted in the complex greenhouse environment, affect the operation of the mechanical driving device, and reduce the operation efficiency.

[0004] At present, there is no path planning method and obstacle avoidance measure for intelligent execution equipment in the greenhouse environment. The existing path planning related technology usually adopts a preset A* algorithm to find the optimal path after modeling, extracting and calculating the internal structure of the greenhouse by a preset algorithm.

[0005] The traditional environment detection method is usually not accurate and has poor planning effect in the face of the complex environment of the greenhouse. In the greenhouse environment in China, the road surface is uneven and has branches. These weak obstacles will be defined as obstacles that need to be avoided in the traditional algorithm. Too frequent obstacle avoidance will cause the machine to repeatedly detour, resulting in low operation efficiency.

[0006] On the other hand, the road surface of the greenhouse is often wet soil or even muddy road. The traditional algorithm does not design an obstacle avoidance measure for this road surface characteristic. If the muddy road is uniformly determined as a drivable road surface without obstacle avoidance treatment, it will cause the mechanical driving device to run difficultly, increase the operation time and reduce the service life of the machine. Therefore, we propose a path planning method for an intelligent fertilizing vehicle suitable for a greenhouse environment. SUMMARY

[0007] (I) Technical problems solved

[0008] In view of the deficiencies of the prior art, the present application provides a path planning method for an intelligent fertilizing vehicle suitable for a greenhouse environment, which combines SLAM mapping technology, binocular stereo vision algorithm, deep learning analysis data, PRM (probabilistic road map) algorithm and A* algorithm, establishes an intelligent obstacle avoidance servo system for the car, and realizes efficient path planning in the greenhouse environment.

[0009] (II) Technical solutions

[0010] In order to achieve the above-mentioned purpose, the application provides the following technical scheme: a path planning method of an intelligent fertilizing vehicle suitable for a greenhouse environment, comprising the following steps:

[0011] S1: the trolley enters a target area, and an information acquisition module starts to operate; a laser radar changes a scanning angle and a scanning plane through circulation to preliminarily establish a three-dimensional point cloud view;

[0012] S2: according to a set maximum fertilizing range at a time, a fertilizing object is processed in a regional manner; regional environmental factors are scored, and weights a1 of a muddy degree, b1 of a pothole degree and c1 of a collision degree are calculated by using an AHP (analytic hierarchy process) analysis;

[0013] S3: through a deep learning algorithm and a binocular stereo vision algorithm, the muddy degree a, the pothole degree b and the collision degree c of a point cloud of a road surface of a fertilizing area required at this time are analyzed, the risk degree d of each point cloud is calculated, and a risk area is determined according to the d value of each road surface;

[0014] S4: a trolley road surface drivable reference threshold value e is preset based on the overall quality of the trolley, the risk value d of the risk area is compared with the driving reference value e, and a passable area and a prohibited area are selected;

[0015] S5: a driving time to a next fertilizing point is calculated in combination with a time of machine fertilization in the area, a fertilizing point corresponding to a shortest path is selected as the next fertilizing point, and a fertilizing sequence is determined;

[0016] S6: an optimal path is calculated based on a PRM (probabilistic roadmap) algorithm and an A* algorithm, and the optimal path is smoothed by using curve fitting;

[0017] S7: the trolley advances according to the optimal path after smoothing, a laser radar and a binocular camera update a coordinate graph in real time, and the above steps are cyclically executed until a task is completed.

[0018] Preferably, the three-dimensional point cloud view in S1 is a point cloud with the laser radar as an origin, and the point cloud is segmented, filtered and completed.

[0019] Preferably, the scoring in S2 is performed by using an independent device that can communicate with the trolley and contains an expert evaluation system.

[0020] Preferably, the calculation formula in S3 is d=a*a1+b*b1+b*b1.

[0021] Preferably, the overall quality of the trolley in S4 includes a trolley mechanical structure, an overall machine quality, a braking performance and a power performance.

[0022] Preferably, the optimal path acquisition method in S6 is as follows:

[0023] First step: use PRM algorithm to construct path network graph;

[0024] Second step: according to the risk degree and the driving ability of the car, calculate the time required for each path;

[0025] Third step: combine the fertilization sequence and use A* algorithm to calculate the optimal path.

[0026] Preferably, when the driving area has a sudden obstacle that causes the car to deviate from the operation path, the system will use A* algorithm to calculate the shortest path back to the original optimal planning, and if the car repeatedly falls into the obstacle group and cannot return to the optimal path, the vehicle will stop and send an alarm to prompt manual reset.

[0027] (Three) beneficial effects

[0028] Compared with the prior art, the present application provides an intelligent fertilization vehicle path planning method suitable for greenhouse environment, which has the following beneficial effects:

[0029] 1. The intelligent fertilization vehicle path planning method suitable for greenhouse environment improves the adaptability of the fertilization vehicle to complex greenhouse road surface environment. The present application uses laser radar and binocular camera to collect environmental information and construct a three-dimensional grid map, and designs a road surface risk degree calculation method based on expert evaluation system and AHP hierarchical analysis, optimizing the obstacle road surface determination mechanism.

[0030] 2. The intelligent fertilization vehicle path planning method suitable for greenhouse environment combines PRM (Probabilistic Roadmap) algorithm and A* algorithm. This combination strategy makes the path planning algorithm not only find the optimal path, but also reduce the computational burden, improve the running efficiency of the car in the greenhouse environment, and reduce the collision to increase the service life. DETAILED DESCRIPTION

[0031] Figure 1 The present application is a method flowchart. CONCRETE EMBODIMENT

[0032] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0033] Please refer to Figure 1 An intelligent fertilization vehicle path planning method suitable for greenhouse environment, comprising the following steps:

[0034] 1) When the trolley enters the target area, the information acquisition module starts to work, the laser radar changes the scanning angle and plane by circulation to obtain the point cloud with the laser radar as the origin, and the point cloud is segmented, filtered and completed to preliminarily establish a three-dimensional point cloud view.

[0035] 2) The fertilization objects are processed in different regions according to the set maximum fertilization range.

[0036] 3) Through the independent device which can communicate with the trolley and contains an expert evaluation system, the importance of the muddy degree, the pit degree and the collision degree is scored according to the local weather condition, the time period when fertilization is carried out, the road condition, the object of fertilization and the structure type of the greenhouse, and the weight a1 of the muddy degree, the weight b1 of the pit degree and the weight c1 of the collision degree are calculated by AHP hierarchical analysis.

[0037] 4) After analyzing the muddy degree a, the pit degree b and the collision degree c of the road point cloud of the fertilization area required at this time by the deep learning algorithm and the binocular stereo vision algorithm, the risk degree d of each point cloud is calculated according to the formula d=a*a1+b*b1+b*b1. The risk area is determined according to the d value of each road section.

[0038] 5) The trolley road drivable reference threshold value e is set in advance according to the parameters such as the mechanical structure of the trolley, the overall quality, the braking performance and the power performance. The risk value d of the risk area is compared with the driving reference value e to screen out the passable area and the prohibited area.

[0039] 6) After initializing the fertilization starting point, the driving time of the starting point relative to each fertilization point is calculated in combination with the required fertilization time of each fertilization point machine in the region, and the fertilization point corresponding to the shortest path is selected as the next fertilization point. The fertilization sequence is determined.

[0040] 7) The path network graph is constructed by using the PRM (probabilistic road map) algorithm, the time required for each path is calculated according to the risk degree and the driving ability of the trolley, and the optimal path is calculated by using the A* algorithm in combination with the fertilization sequence, and the optimal path is smoothed by using curve fitting.

[0041] 8) The control chip controls the trolley to move forward according to the optimal path, and the laser radar and the binocular camera update the coordinate graph in real time.

[0042] 9) The cycle is executed until the work is completed. When the driving area appears a sudden obstacle to cause the trolley to deviate from the work path for emergency obstacle avoidance, the system will calculate the shortest path to return to the original optimal planning by using the A* algorithm. If the trolley repeatedly falls into the obstacle group and cannot return to the optimal path, the vehicle stops and sends an alarm to prompt manual reset.

[0043] The system designs a road obstacle risk value evaluation mechanism based on an expert system and an AHP (analytic hierarchy process) analysis, and realizes comprehensive evaluation of the road obstacle risk by combining binocular vision algorithm and road point cloud data established by deep learning technology. The obstacle avoidance mechanism of the traditional path planning method is effectively improved, so that the fertilizing vehicle can effectively deal with the path planning scene of a muddy path and a plant-shaded road surface in a greenhouse. The reachability, timeliness and safety of path planning of the greenhouse fertilizing vehicle are improved.

[0044] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. An intelligent fertilization vehicle path planning method suitable for a greenhouse environment, characterized in that, Comprise the following steps: S1: The trolley enters the target area, and the information collection module starts to operate. The laser radar changes the scanning angle and scanning plane by circulation to preliminarily establish a three-dimensional point cloud view; S2: According to the set maximum fertilization range at a time, the fertilization object is processed in different regions, the regional environmental factors are scored, and the weights a1 of the muddy degree, b1 of the pothole degree and c1 of the collision degree are calculated by AHP hierarchical analysis; S3: Through the deep learning algorithm and the binocular stereo vision algorithm, the muddy degree a, the pothole degree b and the collision degree c of the road surface point cloud of the required fertilization area at this time are analyzed, and the risk degree d of each point cloud is calculated. The risk area is determined according to the d value of each road surface; S4: Based on the overall quality of the trolley, the trolley road drivable reference threshold e is preset, the risk value d of the risk area is compared with the driving reference value e, and the passable area and the prohibited passable area are selected; S5: The driving time of the next fertilization point is calculated in combination with the time of machine fertilization in the region, and the fertilization point corresponding to the shortest path is selected as the next fertilization point, so as to determine the fertilization sequence; S6: The optimal path is calculated based on the PRM algorithm and the A* algorithm, and the optimal path is smoothed by curve fitting; S7: The trolley advances according to the smoothed optimal path, the laser radar and the binocular camera update the coordinate graph in real time, and the above steps are executed in circulation until the work is completed.

2. The intelligent fertilization vehicle path planning method suitable for a greenhouse environment according to claim 1, characterized in that: The three-dimensional point cloud view in S1 is the point cloud with the laser radar as the origin, and the point cloud is segmented, filtered and completed.

3. The intelligent fertilization vehicle path planning method suitable for greenhouse environment according to claim 1, characterized in that: The scoring in S2 is performed by an independent device that can communicate with the trolley and contains an expert evaluation system.

4. The intelligent fertilization vehicle path planning method suitable for greenhouse environment according to claim 1, characterized in that: The calculation formula in S3 is d=a*a1+b*b1+b*b1.

5. The intelligent fertilization vehicle path planning method suitable for greenhouse environment according to claim 1, characterized in that: The overall quality of the trolley in S4 includes the mechanical structure of the trolley, the overall quality of the machine, the braking performance and the power performance.

6. The intelligent fertilization vehicle path planning method suitable for greenhouse environment according to claim 1, characterized in that: The optimal path acquisition method in S6 is as follows: First step: Construct a path network graph by using the PRM algorithm; Second step: Calculate the time required for each path according to the risk degree and the driving ability of the trolley; Third step: Calculate the optimal path by using the A* algorithm in combination with the fertilization sequence.

7. The intelligent fertilization vehicle path planning method suitable for greenhouse environment according to claim 1, characterized in that: When the driving area appears a sudden obstacle, the trolley deviates from the work path to avoid the obstacle, the system will calculate the shortest path to return to the original optimal planning path by using the A* algorithm. If the trolley repeatedly falls into the obstacle group and cannot return to the optimal path, the vehicle stops and sends an alarm to prompt manual reset.

Images