| Parking space detection is a key technology for realizing automatic parking.The existing parking spaces are mainly designed for regular parking spaces,but there are a lot of planning and irregular parking situations in the actual parking environment.When a vehicle performs parking space detection,the detection accuracy is often low,and manual intervention is required to adjust and detect a variety of parking spaces,and the level of intelligence is low.Therefore,it is necessary to research the problem of parking spaces detection and improve the intelligence level of vehicles in the situation of irregular parking spaces.It is of great value to improve the scene adaptability of parking spaces detection methods.In view of the above problems,the paper uses laser radar as an environmental awareness method,proposes a data clustering algorithm that takes into account the vehicle contour characteristics,and presents a line segment fitting method that is more suitable for vehicle contour extraction.This paper extracts model parameters by establishing parking space model and then establishes a parking space intelligent detection method based on multi-input and multi-rule fuzzy inference system.Finally,this paper completes the design of the intelligent detection system for parking spaces based on 2D laser radar.The main contents include:(1)Lidar data clustering method considering vehicle contour features.Based on the characteristics of the data characteristics when the parking spaces are scanned,this paper proposes an improved self-adaptive threshold clustering method by analyzing the scanning principle of the laser radar,the relative position relationship of the parking spaces,and the outline characteristics of the vehicle,the results of comparative experiments show that the method is superior to the traditional algorithm in both real-time and accuracy.(2)Line segment fitting methods considered merge and over-segmentation issues.Aiming at the problem of over-combination and over-segmentation in the current segment fitting method.By analyzing the characteristics of LT algorithm and IEPF algorithm,this paper proposes the LT-IEPF line segmentation algorithm that combines the advantages of LT algorithm and IEPF algorithm,and improves the threshold selection method in the process of combination at the same time.The comparison experiments show that the algorithm can effectively reduce the line segment fitting error,which is superior to the improved algorithm in the accuracy of the environment perception.(3)An Intelligent Detection Method for Parking Space Based on Multiple Inputs and Multiple Rules Fuzzy Inference System.First,this paper establishes a spatial model for parking spaces and calculates the relevant parameters of the space model.Then,a multi-input and multi-rule fuzzy inference system is applied to the parking space detection environment.At the same time,the spatial model parameters are extracted as input and the parking space detection results are used as the output to detect the parking spaces.Finally,the simulation environment is built by Prescan simulation software,the comparison experiments show that this method has higher recognition accuracy for irregular parking spaces,which is suitable for a variety of parking spaces and are more adaptive and intelligent.Based on the above research results,this paper designs and implements a parking space intelligent detection system based on 2D laser radar.The application results show that the parking space intelligent detection system designed in this paper can accurately detect irregular parking spaces and has a good adaptability to various parking spaces. |
PDF Full Text Request