Research On Blind Road Detection Algorithm Based On Sparse Representation

Blind road detection is the key technology of blinded equipment,which is of great significance to improve the utilization rate of blind road.In recent years,researchers have proposed a variety of blind road segmentation algorithms,but in the face of various types of blind road and complex scenes in the city,the existing segmentation method exposes the disadvantages of poor robustness and narrow application.Therefore,it is of great value to study the automatic segmentation algorithm of blind road and to increase the robustness and universality of the algorithm while ensuring the precision of segmentation.The existing blind road segmentation algorithm is to extract the color or texture features,and then use the threshold segmentation or clustering method to segment.These algorithms are susceptible to blind road types and external environments.Aiming at this problem,this paper proposes a blind road detection algorithm based on sparse representation.The algorithm considers the overall characteristics of blind road,introduces the idea of machine learning,and transforms the problem of dividing the blind road into classifying each small piece in the image.The main work of this paper is as follows:(1)A blind road segmentation algorithm based on significance detection and improved projection dictionary pair is proposed.In this paper,the dictionary pair learning theory is introduced,and its low rank constraint is improved,and a dictionary pair learning algorithm with strong robustness is obtained.Firstly,the saliency detection algorithms are used to coarse locate the blind road region,then the image piece is used as the processing unit,and the dictionary is learned through the robust projective dictionary pair learning proposed.And the coarse-located image is divided into pieces for sparsely reconstruct on the dictionary.Finally,pieces are classified according to the reconstruction error to achieve the purpose of segmentation.(2)The algorithm of blind road boundary detection and inflection points detection is studied.In the research of the blind road boundary detection,the Canny operator is used to detect the blind boundary.In the research of the inflection point detection,this paper proposes a method to detect the inflection point for the blind path.The method first uses the straight line to fit the blind boundary,then the inflection points are found by calculating the curvature of the boundary points.At the end of this paper,the algorithm proposed in this paper is compared with other algorithms.The experimental results show that the algorithm in the blind road segmentation is better than the existing algorithm,both accuracy and universality.