基于双摄像机的立体视觉图像预处理研究

针对立体图像对的亮度差异问题和图像噪声问题研究图像的强化处理,图像预处理主要完成图像对极线校正和图像质量增强.极线校正部分以对极几何为基础研究立体图像对的校正算法.图像质量增强部分主要研究了立体图像对特有的亮度差异问题,并提出了应用直方图均衡化来校正亮度的方法.之后本文对图像噪声进行了深入的分析,并针对立体图像特点使用均值滤波和Gauss滤波平滑图像.在实验中,验证了算法的可行性,取得了良好的应用效果.

Research on Image Preprocessing Based on the Binocular Stereo Vision

This paper researches the strengthening process of image including the brightness difference of stereo image pair and image noise. The image preprocessing contains polar correction of image pair and image enhancement. The polar correction studies correction algorithm of stereo image pair on basis of epipolar geometry. The image enhancement researches the brightness difference of stereo image pair and creates the method of correction of the brightness by using histogram equalization. According to the feature of stereo image pair, this paper analyzes deeply image noise and smooths image by using mean filter and Gaussian filter. The actual experiment verifies the practicability of the algorithm and achieves good effects in many aspects