Complex vehicle road image boundary optimization method

• Main Authors: WEI FEITING, ZHANG JIAJIA, WANG HUIFENG, GUAN LIMIN, MU KENAN, HUANG HE, NI JINGXUE, CHAI CAIPING, YANG WENGUANG
• Format: Patent
• Language: Chinese; English
• Published: 31.12.2019
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention provides a complex vehicle road image boundary optimization method, which realizes classification of image targets in a complex vehicle road environment. Firstly, carrying out training simulation of a large amount of data through a SegNet algorithm model to obtain rough vehicle road target classification features; obtaining an over-segmentation region of the image by using a simple linear iterative clustering algorithm; determining the category of each pixel in each super-pixel region by combining a neural network obtained by a SegNet algorithm, and finally, optimizing the semantic segmentation result by utilizing the accurate boundary recovery capability of a conditional random field, so that the boundary and small-region target error segmentation optimization of the vehicleroad image is realized. The result shows that the method provided by the invention can improve the segmentation precision of the object boundary. 本发明提出了一种复杂车路图像边界优化方法，实现了复杂车路环境下图像目标的分类。首先通过SegNet算法模型进行大量数据的训练仿真得到粗糙的车路目标分类特征，