Research On Lane Detection Based On RHT Method

With the development of Artificial Intelligence,Data mining and Mobile Internet,autonomous vehicle has become a hotspot in the research of vehicle engineering.As an important research target,accurate and rapid identification of lane mark plays an important role for intelligent vehicle and ADAS.Although the research of lane detection has made some progress,there are still some issues need to be solved such as accuracy and real-time performance.Based on the existing research,this paper improved the accuracy and real-time performance through reducing the proportion of noise and data points selection method.Then the proposed method was tested and analyzed with actual driving experiment data.The main work of this paper listed as the followings:(1)Lane edge detection.In order to reduce the interference of the noise and improve the accuracy of the detection results,the image needs to be preprocessed.Based on the analysis and verification of the lane mark images,the region of interest,image filtering and edge detection method are designed.(2)Lane mark modeling and lane detection.According to the regulations of expressway design,in this paper,the straight-line model was adopted.Based on the existing model-based lane detection algorithm,this paper proposes the improvement which is related to straight-line model.(3)Improvement for lane mark detection based on RHT.In this paper,the improvements are expressed in two aspects.Firstly,the improvement based on the filtering processing of lane which based on angle information and means to reduce the proportion of interference points;secondly,based on the centralized random selection of data points which means to improve the problem of the slow speed during the late period.(4)Validation and results analysis.Based on the research results that list above,this paper carried out corresponding verification and analysis combined with the actual road scene image.The result shows that the proposed method in this paper can reduce the misrecognition rate and time consumption.