文件名称:A_simple_method_to_steoro_match
介绍说明--下载内容均来自于网络,请自行研究使用
汽车防撞,技术路径不外:1.雷达测距防撞;2.视差测距防撞。前者,一旦保有量较大,必定遭遇互相干扰问题;后者,以前主要问题是,算法复杂,实时性差。
本文公开了一种新算法(已申请发明专利),主要运算可以借助硬件组合逻辑模块并行执行,可以极大提高视差测距的实时性,满足汽车防撞的要求。-(Background) Stereo matching, requires in two images to identify two pixels to be matched each other, in the field of machine vision, this is a difficult, but central task To complete the task, the key is to strengthen pixels characteristics (status)In order to strengthen the characteristics of pixels , has developed combination of the pixel neighborhood features technology, a variety of window matching technology, and matching prominent features of pixels only, for example, the edge pixels in the image, or to match with the Sift technical characteristics of pixels The window matching Technology, in theory, the larger the window, the more prominent features of pixels to be matched, but the resulting: 1. the computational size is increase 2. the windows across the actual object boundary, so that, the same characteristics pixel in the other image is inexistence,making computing more difficult, so two questions The feature matching, can acquire sparse disparity map only, and mat
本文公开了一种新算法(已申请发明专利),主要运算可以借助硬件组合逻辑模块并行执行,可以极大提高视差测距的实时性,满足汽车防撞的要求。-(Background) Stereo matching, requires in two images to identify two pixels to be matched each other, in the field of machine vision, this is a difficult, but central task To complete the task, the key is to strengthen pixels characteristics (status)In order to strengthen the characteristics of pixels , has developed combination of the pixel neighborhood features technology, a variety of window matching technology, and matching prominent features of pixels only, for example, the edge pixels in the image, or to match with the Sift technical characteristics of pixels The window matching Technology, in theory, the larger the window, the more prominent features of pixels to be matched, but the resulting: 1. the computational size is increase 2. the windows across the actual object boundary, so that, the same characteristics pixel in the other image is inexistence,making computing more difficult, so two questions The feature matching, can acquire sparse disparity map only, and mat
(系统自动生成,下载前可以参看下载内容)
下载文件列表
A_simple_method_to_steoro_match.PDF