关于图像透射变换的一点总结

前段时间为将机器人单目视觉图像从透视图转化为鸟瞰图做了一点研究，这两天为了一个利用图像偏振的野外积水检测，需要将透视图转化为鸟瞰图。上次的这块东西没有好好的总结一下，所以今天花了下午不少时间回顾上次做的研究。ps不做好总结害死人啊，光找上次的那几个工程就花了半天时间，搜索了整个E盘，结果却发现原来那几个工程都在桌面上，浮云啊。

言归正传，这里用到的其实就是学习OpenCV这本书12章ex12里面的东西。无非就是利用事先标定好的摄像头内参数和畸变矩阵将图像去畸变，计算透射变换矩阵，然后将原来的透视图转化为鸟瞰图而已。

#include <cv.h> #include <highgui.h> #include <stdio.h> #include <stdlib.h> void help(){ printf("Birds eye view/n/n" " birdseye board_w board_h intrinsics_mat.xml distortion_mat.xml checker_image /n/n" "Where: board_{w,h} are the # of internal corners in the checkerboard/n" " intrinsic intrinsic path/name of matrix from prior calibration/n" " distortion distortion path/name of matrix from pror calibration/n" " checker_image is the path/name of image of checkerboard on the plane /n" " Frontal view of this will be shown./n/n" " ADJUST VIEW HEIGHT using keys 'u' up, 'd' down. ESC to quit./n/n"); } int main(int argc, char* argv[]) { argv[5]="test"; argc=6; if(argc != 6){ printf("/nERROR: too few parameters/n"); help(); return -1; } help(); //INPUT PARAMETERS: int board_w =9;// atoi("12"); int board_h =6;// atoi("12