opencv-不同方式操作像素效率对比

#include <opencv2/core/core.hpp>#include <opencv2/highgui/highgui.hpp>#include <iostream>using namespace cv;// 下标M.at<float>(i,j)  方法1-1cv::Mat inverseColor1(cv::Mat srcImage){cv::Mat tempImage = srcImage.clone();int row = tempImage.rows;int col = tempImage.cols;// 分别对各个通道进行反色处理for (int i = 0; i < row; i++){for (int j = 0; j < col; j++){tempImage.at<cv::Vec3b>(i, j)[0] = 255 - tempImage.at<cv::Vec3b>(i, j)[0];tempImage.at<cv::Vec3b>(i, j)[1] = 255 - tempImage.at<cv::Vec3b>(i, j)[1];tempImage.at<cv::Vec3b>(i, j)[2] = 255 - tempImage.at<cv::Vec3b>(i, j)[2];}}return tempImage;}// 方法1-2 下标M::at<float>(i,j)  cv::Mat inverseColor2(cv::Mat srcImage){cv::Mat tempImage = srcImage.clone();int row = tempImage.rows;// 图像像素行实际的宽度int step = tempImage.step;// 直接对像素进行反色处理for (int i = 0; i < row; i++){for (int j = 0; j < step; j++){tempImage.at<uchar>(i, j) = 255 - tempImage.at<uchar>(i, j);}}return tempImage;}// 方法2  指针遍历Mat::ptrcv::Mat inverseColor3(cv::Mat srcImage){cv::Mat tempImage = srcImage.clone();int row = tempImage.rows;// 将3通道转换为单通道int nStep = tempImage.cols * tempImage.channels();for(int i = 0; i < row; i++) {// 取源图像的指针const uchar* pSrcData = srcImage.ptr<uchar>(i);// 将输出数据指针存放输出图像uchar* pResultData = tempImage.ptr<uchar>(i);for(int j=0; j < nStep; j++){pResultData[j]= cv::saturate_cast<uchar>(255 - pSrcData[j]);}}return tempImage;}// 方法3 使用迭代器MatConstIteratorcv::Mat inverseColor4(cv::Mat srcImage){cv::Mat tempImage = srcImage.clone();// 初始化源图像迭代器cv::MatConstIterator_<cv::Vec3b> srcIterStart  = srcImage.begin<cv::Vec3b>();cv::MatConstIterator_<cv::Vec3b> srcIterEnd = srcImage.end<cv::Vec3b>();// 初始化输出图像迭代器cv::MatIterator_<cv::Vec3b> resIterStart = tempImage.begin<cv::Vec3b>();cv::MatIterator_<cv::Vec3b> resIterEnd = tempImage.end<cv::Vec3b>();// 遍历图像反色处理while( srcIterStart != srcIterEnd ){(*resIterStart)[0] = 255 - (*srcIterStart)[0];(*resIterStart)[1] = 255 - (*srcIterStart)[1];(*resIterStart)[2] = 255 - (*srcIterStart)[2];// 迭代器递增srcIterStart++;resIterStart++;}return tempImage;}// 方法4 改进的指针方法isContinuouscv::Mat inverseColor5(cv::Mat srcImage){int row = srcImage.rows;int col = srcImage.cols;cv::Mat tempImage = srcImage.clone();// 判断是否是连续图像，即是否有像素填充if( srcImage.isContinuous() && tempImage.isContinuous() ){row = 1;col = col * srcImage.rows * srcImage.channels();}// 遍历图像的每个像素for(int i = 0; i < row; i++){const uchar* pSrcData = srcImage.ptr<uchar>(i);uchar* pResultData = tempImage.ptr<uchar>(i);for(int j = 0; j < col; j++){*pResultData++ = 255 - *pSrcData++;}}return tempImage;}// 方法5 LUT查表法cv::Mat inverseColor6(cv::Mat srcImage){int row = srcImage.rows;int col = srcImage.cols;cv::Mat tempImage = srcImage.clone();// 建立LUT 反色tableuchar LutTable[256];for (int i = 0; i < 256; ++i)LutTable[i] = 255 - i;cv::Mat lookUpTable(1, 256, CV_8U);uchar* pData = lookUpTable.data; // 建立映射表for( int i = 0; i < 256; ++i)pData[i] = LutTable[i];// 应用索引表进行查找cv::LUT(srcImage, lookUpTable, tempImage);return tempImage;}int main(){// 装载图像转为灰度图像cv::Mat srcImage = cv::imread("flower3.jpg");if (!srcImage.data)   return -1;cv::imshow("srcImage", srcImage);cv::Mat resultImg1,resultImg2,resultImg3;cv::Mat resultImg4,resultImg5,resultImg6;// 测试方法1-1double tTime;tTime = (double)getTickCount();const int nTimes = 100;for(int i  = 0;  i < nTimes; i++){resultImg1 = inverseColor1(srcImage);}tTime= 1000*((double)getTickCount() - tTime)/getTickFrequency();tTime /= nTimes;std::cout << "test1-1: "  << tTime << std::endl;// 测试方法1-2tTime = (double)getTickCount();for(int i  = 0;  i < nTimes; i++){resultImg2 = inverseColor2(srcImage);}tTime= 1000*((double)getTickCount() - tTime)/getTickFrequency();tTime /= nTimes;std::cout << "test1-2: "  << tTime << std::endl;// 测试方法2tTime = (double)getTickCount();for(int i  = 0;  i < nTimes; i++){resultImg3 = inverseColor3(srcImage);}tTime= 1000*((double)getTickCount() - tTime)/getTickFrequency();tTime /= nTimes;std::cout << "test2: "  << tTime << std::endl;// 测试方法3tTime = (double)getTickCount();for(int i  = 0;  i < nTimes; i++){resultImg4 = inverseColor4(srcImage);}tTime= 1000*((double)getTickCount() - tTime)/getTickFrequency();tTime /= nTimes;std::cout << "test3: "  << tTime << std::endl;// 测试方法4tTime = (double)getTickCount();for(int i  = 0;  i < nTimes; i++){resultImg5 = inverseColor5(srcImage);}tTime= 1000*((double)getTickCount() - tTime)/getTickFrequency();tTime /= nTimes;std::cout << "test4: "  << tTime << std::endl;// 测试方法5tTime = (double)getTickCount();for(int i  = 0;  i < nTimes; i++){resultImg6 = inverseColor5(srcImage);}tTime= 1000*((double)getTickCount() - tTime)/getTickFrequency();tTime /= nTimes;std::cout << "test5: "  << tTime << std::endl;cv::imshow("resultImg1", resultImg1);cv::imshow("resultImg2", resultImg2);cv::imshow("resultImg3", resultImg3);cv::imshow("resultImg4", resultImg4);cv::imshow("resultImg5", resultImg5);cv::imshow("resultImg6", resultImg6);cv::waitKey(0);return 0;}

http://lib.csdn.net/base/opencv