本文介紹了在 C++ 中旋轉(zhuǎn)圖像而不使用 OpenCV 函數(shù)的處理方法，對(duì)大家解決問(wèn)題具有一定的參考價(jià)值，需要的朋友們下面隨著小編來(lái)一起學(xué)習(xí)吧！
問(wèn)題描述
說(shuō)明:-我試圖在不使用 C++ 中的 OpenCV 函數(shù)的情況下旋轉(zhuǎn)圖像.旋轉(zhuǎn)中心不必是圖像的中心.它可能是一個(gè)不同的點(diǎn)(從圖像中心偏移).到目前為止，我遵循各種來(lái)源進(jìn)行圖像插值，并且我知道
更新:-
在受到與此問(wèn)題相關(guān)的許多答案以及下面最詳盡、最有用和最慷慨的答案的啟發(fā)后，我可以修復(fù)我的 OpenCV 代碼以獲得所需的結(jié)果.
修改后的代碼:
//平凡常量constexpr 雙圓周率 = 3.1415926535897932384626433832795；/*!* rief 函數(shù)生成變換矩陣* param angle 是用戶輸入的旋轉(zhuǎn)角度* param pivot 是 x 軸和 y 軸的平移量*
eturn 平移矩陣*/cv::Mat CreateTransMat(double angle, std::pair<int, int> &pivot) {角度 = Pi * 角度/180；返回(cv::Mat_(3, 3)<(1, 2) <<(0, 0)/trans_mat.at(0, 2),trans_mat.at<double>(0, 1)/trans_mat.at<double>(0, 2));}/*!* rief 基于旋轉(zhuǎn)角度和平移變換圖像的函數(shù)矩陣.當(dāng)旋轉(zhuǎn)和平移同時(shí)發(fā)生時(shí)，兩個(gè)矩陣可以合并* param src 是源圖像* param dest 是目標(biāo)圖像* param trans_mat 是變換(旋轉(zhuǎn)/平移)矩陣*/void ImageTransform(const cv::Mat &src, const cv::Mat &trans_mat, cv::Mat &dest) {int src_rows = src.rows;int src_cols = src.cols;int dest_rows = dest.rows;int dest_cols = dest.cols;const cv::Mat inverse_mat = trans_mat.inv();//#pragma omp parallel for simdfor (int row = 0; row < dest_rows; row++) {//#pragma omp parallel for simdfor (int col = 0; col < dest_cols; col++) {cv::Mat src_pos = CoordTransform(inverse_mat,(cv::Mat_(3, 1) (src_pos.at(0, 0) + 0.5);const int y_actual = static_cast(src_pos.at(0, 1) + 0.5);如果 (x_actual >= 0 && x_actual < src_cols &&y_actual >= 0 &&y_實(shí)際(row, col) = src.at(y_actual, x_actual);別的dest.at(row, col) = cv::Vec3b(0, 0, 0);}}}/*!* rief 命令行參數(shù)輸入的用戶手冊(cè)*/無(wú)效用法(){std::cout <<命令輸入:-
"<<"./ImageTransform <圖像><旋轉(zhuǎn)角度>><<std::endl;}/*!* rief 主函數(shù)讀取圖像的用戶輸入位置，然后應(yīng)用所需的轉(zhuǎn)換(旋轉(zhuǎn)/平移)*/int main(int argc, char *argv[]){自動(dòng)啟動(dòng) = std::chrono::steady_clock::now();if (argc == 0 || argc <3)用法();別的 {雙學(xué)位 = std::stod(argv[2]);雙角 = 度數(shù) * CV_PI/180.；cv::Mat src_img = cv::imread(argv[1]);std::pairnull_trans = std::make_pair(0, 0);std::pair翻譯_初始 =std::make_pair(src_img.cols/2 + 1, src_img.rows/2 + 1);std::pair翻譯_最終 =std::make_pair(0, -src_img.rows/2 - 4);如果(！src_img.data){std::cout <<圖像空"<<std::endl;簡(jiǎn)歷::等待鍵(0);}簡(jiǎn)歷:: imshow(來(lái)源"，src_img)；cv::Mat dest_img = cv::Mat(static_cast(2 * src_img.rows),static_cast(2 * src_img.cols),src_img.type());cv::Mat trans_mat1 = CreateTransMat(degree, translation_initial);ImageTransform(src_img, trans_mat1, dest_img);cv::imshow(臨時(shí)", dest_img);簡(jiǎn)歷::墊中間_img = dest_img;dest_img.release();dest_img = cv::Mat(src_img.rows, src_img.cols, src_img.type());cv::Mat trans_mat2 = CreateTransMat(0, translation_final);ImageTransform(interim_img, trans_mat2, dest_img);cv::imshow(最終圖像", dest_img);簡(jiǎn)歷::等待鍵(10)；}自動(dòng)結(jié)束 = std::chrono::steady_clock::now();自動(dòng)差異 = 結(jié)束 - 開(kāi)始；std::cout <<std::chrono::duration <double, std::milli>(diff).count() <<"毫秒"<<std::endl;} 輸入圖像
旋轉(zhuǎn)圖像
解決方案
首先，我必須承認(rèn)我同意 來(lái)自我最近寫(xiě)的另一個(gè)答案.(已使用 PPM 文件格式，因?yàn)樗枰钌俚奈募?I/O 代碼.)
接下來(lái)，我使用了linMath.h(我用于 3D 轉(zhuǎn)換的最小數(shù)學(xué)集合)為 2D 轉(zhuǎn)換創(chuàng)建最小數(shù)學(xué)集合–linMath.h:
#ifndef LIN_MATH_H#define LIN_MATH_H#include #include <cassert>#include <cmath>extern const double Pi;模板內(nèi)聯(lián)值 degToRad(值角度){返回 (VALUE)Pi * 角度/(VALUE)180;}模板內(nèi)聯(lián)值 radToDeg(VALUE 角度){返回 (VALUE)180 * 角度/(VALUE)Pi;}枚舉 ArgNull { Null };模板struct Vec2T {typedef VALUE 值；值 x, y;//默認(rèn)構(gòu)造函數(shù)(使元素未初始化)Vec2T() { }Vec2T(ArgNull): x((Value)0), y((Value)0) { }Vec2T(值 x, 值 y): x(x), y(y) { }};typedef Vec2Tvec2f;typedef Vec2Tvec2;模板struct Vec3T {typedef VALUE 值；值 x, y, z;//默認(rèn)構(gòu)造函數(shù)(使元素未初始化)Vec3T() { }Vec3T(ArgNull): x((Value)0), y((Value)0), z((Value)0) { }Vec3T(x 值，y 值，z 值):x(x), y(y), z(z) { }Vec3T(const Vec2T &xy, Value z): x(xy.x), y(xy.y), z(z) { }顯式運(yùn)算符 Vec2T() const { return Vec2T(x, y);}const Vec2f xy() const { return Vec2f(x, y);}const Vec2f xz() const { return Vec2f(x, z);}const Vec2f yz() const { return Vec2f(y, z);}};typedef Vec3Tvec3f;typedef Vec3Tvec3;枚舉 ArgInitIdent { InitIdent };枚舉 ArgInitTrans { InitTrans };枚舉 ArgInitRot { InitRot };枚舉 ArgInitScale { InitScale };枚舉 ArgInitFrame { InitFrame };模板結(jié)構(gòu) Mat3x3T {聯(lián)合{價(jià)值補(bǔ)償[3 * 3];結(jié)構(gòu){值_00、_01、_02；值_10、_11、_12；值_20、_21、_22；};};//默認(rèn)構(gòu)造函數(shù)(使元素未初始化)Mat3x3T() { }//構(gòu)造函數(shù)以按元素構(gòu)建矩陣Mat3x3T(價(jià)值_00，價(jià)值_01，價(jià)值_02，價(jià)值_10，價(jià)值_11，價(jià)值_12，價(jià)值_20、價(jià)值_21、價(jià)值_22):_00(_00), _01(_01), _02(_02),_10(_10), _11(_11), _12(_12),_20(_20)、_21(_21)、_22(_22){ }//構(gòu)造單位矩陣的構(gòu)造函數(shù)Mat3x3T(ArgInitIdent):_00((VALUE)1)、_01((VALUE)0)、_02((VALUE)0)、_10((VALUE)0)、_11((VALUE)1)、_12((VALUE)0)、_20((VALUE)0)、_21((VALUE)0)、_22((VALUE)1){ }//構(gòu)造一個(gè)用于翻譯的矩陣Mat3x3T(ArgInitTrans, const Vec2T &t):_00((VALUE)1)、_01((VALUE)0)、_02((VALUE)t.x)、_10((VALUE)0), _11((VALUE)1), _12((VALUE)t.y),_20((VALUE)0)、_21((VALUE)0)、_22((VALUE)1){ }//構(gòu)造函數(shù)來(lái)構(gòu)建旋轉(zhuǎn)矩陣Mat3x3T(ArgInitRot, VALUE 角度):_00(std::cos(angle)), _01(-std::sin(angle)), _02((VALUE)0),_10(std::sin(angle))、_11(std::cos(angle))、_12((VALUE)0)、_20((VALUE)0)、_21((VALUE)0)、_22((VALUE)1){ }//構(gòu)造函數(shù)來(lái)構(gòu)建平移/旋轉(zhuǎn)矩陣Mat3x3T(ArgInitFrame, const Vec2T &t, VALUE 角度):_00(std::cos(angle)), _01(-std::sin(angle)), _02((VALUE)t.x),_10(std::sin(angle))、_11(std::cos(angle))、_12((VALUE)t.y)、_20((VALUE)0)、_21((VALUE)0)、_22((VALUE)1){ }//構(gòu)造函數(shù)以構(gòu)建用于縮放的矩陣Mat3x3T(ArgInitScale, VALUE sx, VALUE sy):_00((VALUE)sx)、_01((VALUE)0)、_02((VALUE)0)、_10((VALUE)0)、_11((VALUE)sy)、_12((VALUE)0)、_20((VALUE)0)、_21((VALUE)0)、_22((VALUE)1){ }//允許使用 [][] 訪問(wèn)的運(yùn)算符VALUE* 運(yùn)算符 [] (int i){斷言(i >= 0 && i < 3);返回補(bǔ)償 + 3 * i;}//允許使用 [][] 訪問(wèn)的運(yùn)算符const VALUE* 運(yùn)算符 [] (int i) const{斷言(i >= 0 && i < 3);返回補(bǔ)償 + 3 * i;}//將矩陣與矩陣相乘 ->矩陣Mat3x3T 運(yùn)算符 * (const Mat3x3T &mat) const{返回 Mat3x3T(_00 * mat._00 + _01 * mat._10 + _02 * mat._20，_00 * mat._01 + _01 * mat._11 + _02 * mat._21，_00 * mat._02 + _01 * mat._12 + _02 * mat._22，_10 * mat._00 + _11 * mat._10 + _12 * mat._20，_10 * mat._01 + _11 * mat._11 + _12 * mat._21，_10 * mat._02 + _11 * mat._12 + _12 * mat._22，_20 * mat._00 + _21 * mat._10 + _22 * mat._20，_20 * mat._01 + _21 * mat._11 + _22 * mat._21，_20 * mat._02 + _21 * mat._12 + _22 * mat._22);}//將矩陣與向量相乘 ->向量Vec3T運(yùn)算符 * (const Vec3T &vec) const{返回 Vec3T(_00 * vec.x + _01 * vec.y + _02 * vec.z，_10 * vec.x + _11 * vec.y + _12 * vec.z，_20 * vec.x + _21 * vec.y + _22 * vec.z)；}};typedef Mat3x3TMat3x3f;typedef Mat3x3T;Mat3x3;模板std::ostream&運(yùn)算符<<(std::ostream &out, const Mat3x3T &m){回來(lái)<<m._00＜＜' ' <<m._01＜＜' ' <<m._02＜＜'
'<<m._10 ＜＜' ' <<m._11 <<' ' <<m._12＜＜'
'<<m._20＜＜' ' <<m._21＜＜' ' <<m._22＜＜'
';}/* 計(jì)算矩陣的行列式.** det = |M|** mat ... 矩陣*/模板值行列式(const Mat3x3T&mat){返回 mat._00 * mat._11 * mat._22+ mat._01 * mat._12 * mat._20+ mat._02 * mat._10 * mat._21- mat._20 * mat._11 * mat._02- mat._21 * mat._12 * mat._00- mat._22 * mat._10 * mat._01;}/* 返回正則矩陣的逆矩陣.** mat 矩陣反轉(zhuǎn)* eps epsilon 矩陣的規(guī)律性*/模板Mat3x3T倒置(const Mat3x3T&mat, VALUE eps = (VALUE)1E-10){斷言(eps >=(值)0)；//計(jì)算行列式并檢查它是否不等于 0//(否則，矩陣是奇異的！)常量值 det = 行列式(墊)；if (std::abs(det) (detInvPos * (mat._11 * mat._22 - mat._12 * mat._21),detInvNeg * (mat._01 * mat._22 - mat._02 * mat._21),detInvPos * (mat._01 * mat._12 - mat._02 * mat._11),detInvNeg * (mat._10 * mat._22 - mat._12 * mat._20),detInvPos * (mat._00 * mat._22 - mat._02 * mat._20),detInvNeg * (mat._00 * mat._12 - mat._02 * mat._10),detInvPos * (mat._10 * mat._21 - mat._11 * mat._20),detInvNeg * (mat._00 * mat._21 - mat._01 * mat._20),detInvPos * (mat._00 * mat._11 - mat._01 * mat._10));}#endif//LIN_MATH_H 以及linMath.cc中Pi的定義:
#include "linmath.h"const double Pi = 3.1415926535897932384626433832795;有了所有可用的工具，我制作了示例應(yīng)用程序 xformRGBImg.cc:
#include #include <fstream>#include #include <字符串>#include "linMath.h"#include "image.h"#include "imagePPM.h"typedef unsigned int uint;結(jié)構(gòu)錯(cuò)誤{const std::string 文本；錯(cuò)誤(常量字符*文本):文本(文本){}};const char* readArg(int &i, int argc, char **argv){++i;if (i >= argc) throw Error("缺少參數(shù)！");返回 argv[i];}uint readArgUInt(int &i, int argc, char **argv){const char *arg = readArg(i, argc, argv);字符 * 結(jié)束；const unsigned long value = strtoul(arg, &end, 0);if (arg == end || *end) throw Error("應(yīng)為無(wú)符號(hào)整數(shù)值！");if ((uint)value != value) throw Error("無(wú)符號(hào)整數(shù)溢出！");返回(單位)值；}double readArgDouble(int &i, int argc, char **argv){const char *arg = readArg(i, argc, argv);字符 * 結(jié)束；const double value = strtod(arg, &end);if (arg == end || *end) throw Error("需要浮點(diǎn)值！");返回值；}std::pair調(diào)整大小(int &i，int argc，char **argv){const uint w = readArgUInt(i, argc, argv);const uint h = readArgUInt(i, argc, argv);返回 std::make_pair(w, h);}Mat3x3 翻譯(int &i，int argc，char **argv){const double x = readArgDouble(i, argc, argv);const double y = readArgDouble(i, argc, argv);返回 Mat3x3(InitTrans, Vec2(x, y));}Mat3x3 旋轉(zhuǎn)(int &i，int argc，char **argv){const double angle = readArgDouble(i, argc, argv);返回 Mat3x3(InitRot, degToRad(angle));}Mat3x3 比例(int &i，int argc，char **argv){const double x = readArgDouble(i, argc, argv);const double y = readArgDouble(i, argc, argv);返回 Mat3x3(InitScale, x, y);}Vec2 變換(const Mat3x3 &mat，const Vec2 &pos){const Vec3 pos_ = mat * Vec3(pos, 1.0);返回 Vec2(pos_.x/pos_.z, pos_.y/pos_.z);}空變換(const Image &imgSrc, const Mat3x3 &mat, Image &imgDst,int rgbFail = 0x808080){const Mat3x3 matInv = invert(mat);for (int y = 0; y < imgDst.h(); ++y) {for (int x = 0; x 大小輸出(0, 0);Mat3x3 mat(InitIdent);for (int i = 3; i < argc; ++i) 試試 {const std::string cmd = argv[i];if (cmd == "resize") sizeOut = resize(i, argc, argv);else if (cmd == "translate") mat = translate(i, argc, argv) * mat;else if (cmd == "rotate") mat = rotate(i, argc, argv) * mat;else if (cmd == "scale") mat = scale(i, argc, argv) * mat;別的 {std::cerr <<"錯(cuò)誤的命令！
";std::cout <<用法;返回 1；}} catch (const Error &error) {std::cerr <<$ 處的錯(cuò)誤參數(shù)" <<我＜＜"
"<<錯(cuò)誤文本<<'
';std::cout <<用法;返回 1；}//讀取圖像圖片 imgSrc;{ std::ifstream fIn(inFile.c_str(), std::ios::binary);如果(！readPPM(fIn，imgSrc)){std::cerr <<閱讀"<<文件中<<"'失??！
";返回 1；}}//設(shè)置輸出圖像大小如果(sizeOut.first * sizeOut.second == 0){sizeOut = std::make_pair(imgSrc.w(), imgSrc.h());}//變換圖像圖像 imgDst;imgDst.resize(sizeOut.first, sizeOut.second, 3 * sizeOut.second);變換(imgSrc，墊，imgDst)；//寫(xiě)入圖像{ std::ofstream fOut(outFile.c_str(), std::ios::binary);if (!writePPM(fOut, imgDst) || (fOut.close(), !fOut.good())) {std::cerr <<寫(xiě)'"<<輸出文件<<"'失??！
";返回 1；}}//完畢返回0；} 注意:
命令行參數(shù)按順序處理.每個(gè)轉(zhuǎn)換命令從左乘到已經(jīng)組合的轉(zhuǎn)換矩陣，從一個(gè)單位矩陣開(kāi)始.這是因?yàn)樽儞Q的串聯(lián)導(dǎo)致矩陣的逆序乘法.(矩陣乘法是右結(jié)合的.)
例如變換的對(duì)應(yīng)矩陣:
x' = 翻譯(x)
x" = 旋轉(zhuǎn)(x')
x"' = 比例(x")
這是
x"' = 縮放(旋轉(zhuǎn)(翻譯(x)))
是
Mtransform = Mscale ·M旋轉(zhuǎn) ·M翻譯
和
x"' = Mscale · M旋轉(zhuǎn) ·Mtranslate · x = Mtransform · x
在
尺寸為 300 ×300.
注意:
所有嵌入的圖像都從 PPM 轉(zhuǎn)換為 JPEG(再次在
看起來(lái)像原來(lái)的–身份轉(zhuǎn)換應(yīng)該是什么.
現(xiàn)在，旋轉(zhuǎn) 30°:
$ ./xformRGBImg cat.ppm cat.rot30.ppm 旋轉(zhuǎn) 30$要繞某個(gè)中心旋轉(zhuǎn)，有一個(gè)相應(yīng)的方法.需要前后翻譯:
$ ./xformRGBImg cat.ppm cat.rot30c150,150.ppm 平移 -150 -150 旋轉(zhuǎn) 30 平移 150 150$輸出圖像可以用 w · 調(diào)整大小√2 ×·√2 以適應(yīng)任何中心旋轉(zhuǎn).
因此，輸出圖像的大小調(diào)整為 425 ×425 其中最后一次翻譯分別調(diào)整為translate 212.5 212.5:
$ ./xformRGBImg cat.ppm cat.rot30c150,150.425x425.ppm 調(diào)整大小 425 425 平移 -150 -150 旋轉(zhuǎn) 30 平移 212.5 212.5$尚未檢查縮放比例:
$ ./xformRGBImg cat.ppm cat.rot30c150,150s0.7,0.7.ppm 平移 -150 -150 旋轉(zhuǎn) 30 縮放 0.7 0.7 平移 150 150$最后，公平地說(shuō)，我想提一下大哥".我的小玩具工具:ImageMagick.
Description :- I am trying to rotate an image without using OpenCV functions in C++. The rotation center need not be the center of the image. It could be a different point (offset from the image center). So far I followed a variety of sources to do image interpolation and I am aware of a source which does the job perfectly in MATLAB. I tried to mimic the same in C++ without OpenCV functions. But I am not getting the expected rotated image. Instead my output appears like a small horizontal line on the screen.
void RotateNearestNeighbor(cv::Mat src, double angle) {
int oldHeight = src.rows;
int oldWidth = src.cols;
int newHeight = std::sqrt(2) * oldHeight;
int newWidth = std::sqrt(2) * oldWidth;
cv::Mat output = cv::Mat(newHeight, newWidth, src.type());
double ctheta = cos(angle);
double stheta = sin(angle);
for (size_t i = 0; i < newHeight; i++) {
for (size_t j = 0; j < newWidth; j++) {
int oldRow = static_cast<int> ((i - newHeight / 2) * ctheta +
(j - newWidth / 2) * stheta + oldHeight / 2);
int oldCol = static_cast<int> (-(i - newHeight / 2) * stheta +
(j - newWidth / 2) * ctheta + oldWidth / 2);
if (oldRow > 0 && oldCol > 0 && oldRow <= oldHeight && oldCol <= oldWidth)
output.at<cv::Vec3b>(i, j) = src.at<cv::Vec3b>(oldRow, oldCol);
else
output.at<cv::Vec3b>(i, j) = cv::Vec3b(0, 0, 0);
}
}
cv::imshow("Rotated cat", output);
}
The following are my input (left side) and output (right side) images
UPDATE : -
After being inspired by many answers related to this question and also the most elaborate, helpful and generous answer below, I could fix my OpenCV code to get the desired result.
Modified Code :
// Trivial constant
constexpr double Pi = 3.1415926535897932384626433832795;
/*!
* rief Function to generate transformation matrix
* param angle is the angle of rotation from user input
* param pivot is the amount of translation in x and y axes
*
eturn translation matrix
*/
cv::Mat CreateTransMat(double angle, std::pair<int, int> &pivot) {
angle = Pi * angle / 180;
return (cv::Mat_<double>(3, 3) << cos(angle), -sin(angle), pivot.first,
sin(angle), cos(angle), pivot.second, 0, 0, 1);
}
/*!
* rief Function to apply coordinate transform from destination to source
* param inv_mat being the inverse transformation matrix for the transform needed
*
eturn pos being the homogeneous coordinates for transformation
*/
cv::Mat CoordTransform(const cv::Mat &inv_mat, const cv::Mat &pos) {
assert(inv_mat.cols == pos.rows);
cv::Mat trans_mat = inv_mat * pos;
return (cv::Mat_<double>(1, 2) <<
trans_mat.at<double>(0, 0) / trans_mat.at<double>(0, 2),
trans_mat.at<double>(0, 1) / trans_mat.at<double>(0, 2));
}
/*!
* rief Function to transform an image based on a rotation angle and translation
matrix. When rotation and translation happen at the same time, the
two matrices can be combined
* param src being source image
* param dest being destination image
* param trans_mat being the transformation (rotation/ translation) matrix
*/
void ImageTransform(const cv::Mat &src, const cv::Mat &trans_mat, cv::Mat &dest) {
int src_rows = src.rows;
int src_cols = src.cols;
int dest_rows = dest.rows;
int dest_cols = dest.cols;
const cv::Mat inverse_mat = trans_mat.inv();
//#pragma omp parallel for simd
for (int row = 0; row < dest_rows; row++) {
//#pragma omp parallel for simd
for (int col = 0; col < dest_cols; col++) {
cv::Mat src_pos = CoordTransform(inverse_mat,
(cv::Mat_<double>(3, 1) << col, row, 1));
const int x_actual = static_cast<int>(src_pos.at<double>(0, 0) + 0.5);
const int y_actual = static_cast<int>(src_pos.at<double>(0, 1) + 0.5);
if (x_actual >= 0 && x_actual < src_cols &&
y_actual >= 0 && y_actual < src_rows)
dest.at<cv::Vec3b>(row, col) = src.at<cv::Vec3b>(y_actual, x_actual);
else
dest.at<cv::Vec3b>(row, col) = cv::Vec3b(0, 0, 0);
}
}
}
/*!
* rief User manual for command-line args input
*/
void Usage() {
std::cout << "COMMAND INPUT : -
" <<
" ./ImageTransform <image> <rotation-angle>" <<
std::endl;
}
/*!
* rief main function to read a user input location for an image and then apply the
required transformations (rotation / translation)
*/
int main(int argc, char *argv[])
{
auto start = std::chrono::steady_clock::now();
if (argc == 0 || argc < 3)
Usage();
else {
double degree = std::stod(argv[2]);
double angle = degree * CV_PI / 180.;
cv::Mat src_img = cv::imread(argv[1]);
std::pair<int, int> null_trans = std::make_pair(0, 0);
std::pair<int, int> translation_initial =
std::make_pair(src_img.cols / 2 + 1, src_img.rows / 2 + 1);
std::pair<int, int> translation_final =
std::make_pair(0, -src_img.rows / 2 - 4);
if (!src_img.data)
{
std::cout << "image null" << std::endl;
cv::waitKey(0);
}
cv::imshow("Source", src_img);
cv::Mat dest_img = cv::Mat(static_cast<int>(2 * src_img.rows),
static_cast<int>(2 * src_img.cols),
src_img.type());
cv::Mat trans_mat1 = CreateTransMat(degree, translation_initial);
ImageTransform(src_img, trans_mat1, dest_img);
cv::imshow("Interim", dest_img);
cv::Mat interim_img = dest_img;
dest_img.release();
dest_img = cv::Mat(src_img.rows, src_img.cols, src_img.type());
cv::Mat trans_mat2 = CreateTransMat(0, translation_final);
ImageTransform(interim_img, trans_mat2, dest_img);
cv::imshow("Final image", dest_img);
cv::waitKey(10);
}
auto end = std::chrono::steady_clock::now();
auto diff = end - start;
std::cout << std::chrono::duration <double, std::milli> (diff).count() <<
" ms" << std::endl;
}
Input image
Rotated image
解決方案
First, I have to admit I agree with generic_opto_guy:
The approach with the loop looks good, so we would need to check the math. On thing I noticed: if (oldRow > 0 && oldCol > 0 && oldRow <= oldHeight && oldCol <= oldWidth) implies you start indexing with 1. I belife that opencv starts indexing with 0.
For all that, I couldn't resist to answer. (May be, it's just an image phase of mine.)
Instead of fiddling with sin() and cos(), I would recommend to use matrix transformation. At the first glance, this might appear over-engineered but later you will recognize that it bears much more flexibility. With a transformation matrix, you can express a lot of transformations (translation, rotation, scaling, shearing, projection) as well as combining multiple transformations into one matrix.
(A teaser for what is possible: SO: How to paint / deform a QImage in 2D?)
In an image, the pixels may be addressed by 2d coordinates. Hence a 2×2 matrix comes into mind but a 2×2 matrix cannot express translations. For this, homogeneous coordinates has been introduced – a math trick to handle positions and directions in the same space by extending the dimension by one.
To make it short, a 2d position (x, y) has the homogeneous coordinates (x, y, 1).
A position transformed with a transformation matrix:
v′ = M · v.
This may or may not change the value of third component. To convert the homogeneous coordinate to 2D position again, x and y has to be divided by 3rd component.
Vec2 transform(const Mat3x3 &mat, const Vec2 &pos)
{
const Vec3 pos_ = mat * Vec3(pos, 1.0);
return Vec2(pos_.x / pos_.z, pos_.y / pos_.z);
}
To transform a source image into a destination image, the following function can be used:
void transform(
const Image &imgSrc, const Mat3x3 &mat, Image &imgDst,
int rgbFail = 0x808080)
{
const Mat3x3 matInv = invert(mat);
for (int y = 0; y < imgDst.h(); ++y) {
for (int x = 0; x < imgDst.w(); ++x) {
const Vec2 pos = transform(matInv, Vec2(x, y));
const int xSrc = (int)(pos.x + 0.5), ySrc = http://pic.html5code.net(int)(pos.y + 0.5);
imgDst.setPixel(x, y,
xSrc >= 0 && xSrc < imgSrc.w() && ySrc >= 0 && ySrc < imgSrc.h()
? imgSrc.getPixel(xSrc, ySrc)
: rgbFail);
}
}
}
Note:
The transformation matrix mat describes the transformation of source image coordinates to destination image coordinates. The nested loops iterate over destination image. Hence, the inverse matrix (representing the reverse transformation) has to be used to get the corresponding source image coordinates which map to the current destination coordinates.
… and the matrix constructor for the rotation:
enum ArgInitRot { InitRot };
template <typename VALUE>
struct Mat3x3T {
union {
VALUE comp[3 * 3];
struct {
VALUE _00, _01, _02;
VALUE _10, _11, _12;
VALUE _20, _21, _22;
};
};
// constructor to build a matrix for rotation
Mat3x3T(ArgInitRot, VALUE angle):
_00(std::cos(angle)), _01(-std::sin(angle)), _02((VALUE)0),
_10(std::sin(angle)), _11( std::cos(angle)), _12((VALUE)0),
_20( (VALUE)0), _21( (VALUE)0), _22((VALUE)1)
{ }
can be used to construct a rotation with angle (in degree):
Mat3x3T<double> mat(InitRot, degToRad(30.0));
Note:
I would like to emphasize how the transformed coordinates are used:
const Vec2 pos = transform(matInv, Vec2(x, y));
const int xSrc = (int)(pos.x + 0.5), ySrc = http://pic.html5code.net(int)(pos.y + 0.5);
Rounding the results to yield one discrete pixel position is actually what is called Nearest Neighbour. Alternatively, the now discarded fractional parts could be used for a linear interpolation between neighbour pixels.
To make a small sample, I first copied image.h, image.cc, imagePPM.h, and imagePPM.cc from another answer I wrote recently. (The PPM file format has been used as it needs minimal code for file I/O.)
Next, I used linMath.h (my minimal math collection for 3D transformations) to make a minimal math collection for 2D transformations – linMath.h:
#ifndef LIN_MATH_H
#define LIN_MATH_H
#include <iostream>
#include <cassert>
#include <cmath>
extern const double Pi;
template <typename VALUE>
inline VALUE degToRad(VALUE angle)
{
return (VALUE)Pi * angle / (VALUE)180;
}
template <typename VALUE>
inline VALUE radToDeg(VALUE angle)
{
return (VALUE)180 * angle / (VALUE)Pi;
}
enum ArgNull { Null };
template <typename VALUE>
struct Vec2T {
typedef VALUE Value;
Value x, y;
// default constructor (leaving elements uninitialized)
Vec2T() { }
Vec2T(ArgNull): x((Value)0), y((Value)0) { }
Vec2T(Value x, Value y): x(x), y(y) { }
};
typedef Vec2T<float> Vec2f;
typedef Vec2T<double> Vec2;
template <typename VALUE>
struct Vec3T {
typedef VALUE Value;
Value x, y, z;
// default constructor (leaving elements uninitialized)
Vec3T() { }
Vec3T(ArgNull): x((Value)0), y((Value)0), z((Value)0) { }
Vec3T(Value x, Value y, Value z): x(x), y(y), z(z) { }
Vec3T(const Vec2T<Value> &xy, Value z): x(xy.x), y(xy.y), z(z) { }
explicit operator Vec2T<Value>() const { return Vec2T<Value>(x, y); }
const Vec2f xy() const { return Vec2f(x, y); }
const Vec2f xz() const { return Vec2f(x, z); }
const Vec2f yz() const { return Vec2f(y, z); }
};
typedef Vec3T<float> Vec3f;
typedef Vec3T<double> Vec3;
enum ArgInitIdent { InitIdent };
enum ArgInitTrans { InitTrans };
enum ArgInitRot { InitRot };
enum ArgInitScale { InitScale };
enum ArgInitFrame { InitFrame };
template <typename VALUE>
struct Mat3x3T {
union {
VALUE comp[3 * 3];
struct {
VALUE _00, _01, _02;
VALUE _10, _11, _12;
VALUE _20, _21, _22;
};
};
// default constructor (leaving elements uninitialized)
Mat3x3T() { }
// constructor to build a matrix by elements
Mat3x3T(
VALUE _00, VALUE _01, VALUE _02,
VALUE _10, VALUE _11, VALUE _12,
VALUE _20, VALUE _21, VALUE _22):
_00(_00), _01(_01), _02(_02),
_10(_10), _11(_11), _12(_12),
_20(_20), _21(_21), _22(_22)
{ }
// constructor to build an identity matrix
Mat3x3T(ArgInitIdent):
_00((VALUE)1), _01((VALUE)0), _02((VALUE)0),
_10((VALUE)0), _11((VALUE)1), _12((VALUE)0),
_20((VALUE)0), _21((VALUE)0), _22((VALUE)1)
{ }
// constructor to build a matrix for translation
Mat3x3T(ArgInitTrans, const Vec2T<VALUE> &t):
_00((VALUE)1), _01((VALUE)0), _02((VALUE)t.x),
_10((VALUE)0), _11((VALUE)1), _12((VALUE)t.y),
_20((VALUE)0), _21((VALUE)0), _22((VALUE)1)
{ }
// constructor to build a matrix for rotation
Mat3x3T(ArgInitRot, VALUE angle):
_00(std::cos(angle)), _01(-std::sin(angle)), _02((VALUE)0),
_10(std::sin(angle)), _11( std::cos(angle)), _12((VALUE)0),
_20( (VALUE)0), _21( (VALUE)0), _22((VALUE)1)
{ }
// constructor to build a matrix for translation/rotation
Mat3x3T(ArgInitFrame, const Vec2T<VALUE> &t, VALUE angle):
_00(std::cos(angle)), _01(-std::sin(angle)), _02((VALUE)t.x),
_10(std::sin(angle)), _11( std::cos(angle)), _12((VALUE)t.y),
_20( (VALUE)0), _21( (VALUE)0), _22((VALUE)1)
{ }
// constructor to build a matrix for scaling
Mat3x3T(ArgInitScale, VALUE sx, VALUE sy):
_00((VALUE)sx), _01( (VALUE)0), _02((VALUE)0),
_10( (VALUE)0), _11((VALUE)sy), _12((VALUE)0),
_20( (VALUE)0), _21( (VALUE)0), _22((VALUE)1)
{ }
// operator to allow access with [][]
VALUE* operator [] (int i)
{
assert(i >= 0 && i < 3);
return comp + 3 * i;
}
// operator to allow access with [][]
const VALUE* operator [] (int i) const
{
assert(i >= 0 && i < 3);
return comp + 3 * i;
}
// multiply matrix with matrix -> matrix
Mat3x3T operator * (const Mat3x3T &mat) const
{
return Mat3x3T(
_00 * mat._00 + _01 * mat._10 + _02 * mat._20,
_00 * mat._01 + _01 * mat._11 + _02 * mat._21,
_00 * mat._02 + _01 * mat._12 + _02 * mat._22,
_10 * mat._00 + _11 * mat._10 + _12 * mat._20,
_10 * mat._01 + _11 * mat._11 + _12 * mat._21,
_10 * mat._02 + _11 * mat._12 + _12 * mat._22,
_20 * mat._00 + _21 * mat._10 + _22 * mat._20,
_20 * mat._01 + _21 * mat._11 + _22 * mat._21,
_20 * mat._02 + _21 * mat._12 + _22 * mat._22);
}
// multiply matrix with vector -> vector
Vec3T<VALUE> operator * (const Vec3T<VALUE> &vec) const
{
return Vec3T<VALUE>(
_00 * vec.x + _01 * vec.y + _02 * vec.z,
_10 * vec.x + _11 * vec.y + _12 * vec.z,
_20 * vec.x + _21 * vec.y + _22 * vec.z);
}
};
typedef Mat3x3T<float> Mat3x3f;
typedef Mat3x3T<double> Mat3x3;
template <typename VALUE>
std::ostream& operator<<(std::ostream &out, const Mat3x3T<VALUE> &m)
{
return out
<< m._00 << ' ' << m._01 << ' ' << m._02 << '
'
<< m._10 << ' ' << m._11 << ' ' << m._12 << '
'
<< m._20 << ' ' << m._21 << ' ' << m._22 << '
';
}
/* computes determinant of a matrix.
*
* det = |M|
*
* mat ... the matrix
*/
template <typename VALUE>
VALUE determinant(const Mat3x3T<VALUE> &mat)
{
return mat._00 * mat._11 * mat._22
+ mat._01 * mat._12 * mat._20
+ mat._02 * mat._10 * mat._21
- mat._20 * mat._11 * mat._02
- mat._21 * mat._12 * mat._00
- mat._22 * mat._10 * mat._01;
}
/* returns the inverse of a regular matrix.
*
* mat matrix to invert
* eps epsilon for regularity of matrix
*/
template <typename VALUE>
Mat3x3T<VALUE> invert(
const Mat3x3T<VALUE> &mat, VALUE eps = (VALUE)1E-10)
{
assert(eps >= (VALUE)0);
// compute determinant and check that it its unequal to 0
// (Otherwise, matrix is singular!)
const VALUE det = determinant(mat);
if (std::abs(det) < eps) throw std::domain_error("Singular matrix!");
// reciproke of determinant
const VALUE detInvPos = (VALUE)1 / det, detInvNeg = -detInvPos;
// compute each element by determinant of sub-matrix which is build
// striking out row and column of pivot element itself
// BTW, the determinant is multiplied with -1 when sum of row and column
// index is odd (chess board rule)
// (This is usually called cofactor of related element.)
// transpose matrix and multiply with 1/determinant of original matrix
return Mat3x3T<VALUE>(
detInvPos * (mat._11 * mat._22 - mat._12 * mat._21),
detInvNeg * (mat._01 * mat._22 - mat._02 * mat._21),
detInvPos * (mat._01 * mat._12 - mat._02 * mat._11),
detInvNeg * (mat._10 * mat._22 - mat._12 * mat._20),
detInvPos * (mat._00 * mat._22 - mat._02 * mat._20),
detInvNeg * (mat._00 * mat._12 - mat._02 * mat._10),
detInvPos * (mat._10 * mat._21 - mat._11 * mat._20),
detInvNeg * (mat._00 * mat._21 - mat._01 * mat._20),
detInvPos * (mat._00 * mat._11 - mat._01 * mat._10));
}
#endif // LIN_MATH_H
and the definition of Pi in linMath.cc:
#include "linmath.h"
const double Pi = 3.1415926535897932384626433832795;
Having all tools available, I made the sample application xformRGBImg.cc:
#include <iostream>
#include <fstream>
#include <sstream>
#include <string>
#include "linMath.h"
#include "image.h"
#include "imagePPM.h"
typedef unsigned int uint;
struct Error {
const std::string text;
Error(const char *text): text(text) { }
};
const char* readArg(int &i, int argc, char **argv)
{
++i;
if (i >= argc) throw Error("Missing argument!");
return argv[i];
}
uint readArgUInt(int &i, int argc, char **argv)
{
const char *arg = readArg(i, argc, argv); char *end;
const unsigned long value = strtoul(arg, &end, 0);
if (arg == end || *end) throw Error("Unsigned integer value expected!");
if ((uint)value != value) throw Error("Unsigned integer overflow!");
return (uint)value;
}
double readArgDouble(int &i, int argc, char **argv)
{
const char *arg = readArg(i, argc, argv); char *end;
const double value = strtod(arg, &end);
if (arg == end || *end) throw Error("Floating point value expected!");
return value;
}
std::pair<uint, uint> resize(int &i, int argc, char **argv)
{
const uint w = readArgUInt(i, argc, argv);
const uint h = readArgUInt(i, argc, argv);
return std::make_pair(w, h);
}
Mat3x3 translate(int &i, int argc, char **argv)
{
const double x = readArgDouble(i, argc, argv);
const double y = readArgDouble(i, argc, argv);
return Mat3x3(InitTrans, Vec2(x, y));
}
Mat3x3 rotate(int &i, int argc, char **argv)
{
const double angle = readArgDouble(i, argc, argv);
return Mat3x3(InitRot, degToRad(angle));
}
Mat3x3 scale(int &i, int argc, char **argv)
{
const double x = readArgDouble(i, argc, argv);
const double y = readArgDouble(i, argc, argv);
return Mat3x3(InitScale, x, y);
}
Vec2 transform(const Mat3x3 &mat, const Vec2 &pos)
{
const Vec3 pos_ = mat * Vec3(pos, 1.0);
return Vec2(pos_.x / pos_.z, pos_.y / pos_.z);
}
void transform(
const Image &imgSrc, const Mat3x3 &mat, Image &imgDst,
int rgbFail = 0x808080)
{
const Mat3x3 matInv = invert(mat);
for (int y = 0; y < imgDst.h(); ++y) {
for (int x = 0; x < imgDst.w(); ++x) {
const Vec2 pos = transform(matInv, Vec2(x, y));
const int xSrc = (int)(pos.x + 0.5), ySrc = http://pic.html5code.net(int)(pos.y + 0.5);
imgDst.setPixel(x, y,
xSrc >= 0 && xSrc < imgSrc.w() && ySrc >= 0 && ySrc < imgSrc.h()
? imgSrc.getPixel(xSrc, ySrc)
: rgbFail);
}
}
}
const char *const usage =
"Usage:
"
" xformRGBImg IN_FILE OUT_FILE [[CMD]...]
"
"
"
"Commands:
"
" resize W H
"
" translate X Y
"
" rotate ANGLE
"
" scale SX SY
";
int main(int argc, char **argv)
{
// read command line arguments
if (argc <= 2) {
std::cerr << "Missing arguments!
";
std::cout << usage;
return 1;
}
const std::string inFile = argv[1];
const std::string outFile = argv[2];
std::pair<uint, uint> sizeOut(0, 0);
Mat3x3 mat(InitIdent);
for (int i = 3; i < argc; ++i) try {
const std::string cmd = argv[i];
if (cmd == "resize") sizeOut = resize(i, argc, argv);
else if (cmd == "translate") mat = translate(i, argc, argv) * mat;
else if (cmd == "rotate") mat = rotate(i, argc, argv) * mat;
else if (cmd == "scale") mat = scale(i, argc, argv) * mat;
else {
std::cerr << "Wrong command!
";
std::cout << usage;
return 1;
}
} catch (const Error &error) {
std::cerr << "Wrong argument at $" << i << "
"
<< error.text << '
';
std::cout << usage;
return 1;
}
// read image
Image imgSrc;
{ std::ifstream fIn(inFile.c_str(), std::ios::binary);
if (!readPPM(fIn, imgSrc)) {
std::cerr << "Reading '" << inFile << "' failed!
";
return 1;
}
}
// set output image size
if (sizeOut.first * sizeOut.second == 0) {
sizeOut = std::make_pair(imgSrc.w(), imgSrc.h());
}
// transform image
Image imgDst;
imgDst.resize(sizeOut.first, sizeOut.second, 3 * sizeOut.second);
transform(imgSrc, mat, imgDst);
// write image
{ std::ofstream fOut(outFile.c_str(), std::ios::binary);
if (!writePPM(fOut, imgDst) || (fOut.close(), !fOut.good())) {
std::cerr << "Writing '" << outFile << "' failed!
";
return 1;
}
}
// done
return 0;
}
Note:
The command line arguments are processed in order. Each transformation command is multiplied from left to the already combined transformation matrix, starting with an identity matrix. This is because a concatenation of transformations results in the reverse ordered multiplication of matrices. (The matrix multiplication is right associative.)
E.g. the corresponding matrix for a transform:
x' = translate(x)
x" = rotate(x')
x"' = scale(x")
which is
x"' = scale(rotate(translate(x)))
is
Mtransform = Mscale · Mrotate · Mtranslate
and
x"' = Mscale · Mrotate · Mtranslate · x = Mtransform · x
Compiled and tested in cygwin:
$ g++ -std=c++11 -o xformRGBImg image.cc imagePPM.cc linMath.cc xformRGBImg.cc
$ ./xformRGBImg
Missing arguments!
Usage:
xformRGBImg IN_FILE OUT_FILE [[CMD]...]
Commands:
resize W H
translate X Y
rotate ANGLE
scale SX SY
$
Finally, a sample image cat.jpg (converted to PPM in GIMP):
with size 300 × 300.
Note:
All embedded images are converted from PPM to JPEG (in GIMP again). (PPM is not supported in image upload, nor can I imagine that any browser can display it properly.)
To start with a minimum:
$ ./xformRGBImg cat.ppm cat.copy.ppm
$
It looks like the original – what should be expected by an identity transform.
Now, a rotation with 30°:
$ ./xformRGBImg cat.ppm cat.rot30.ppm rotate 30
$
To rotate about a certain center, there is a resp. translation before and afterwards needed:
$ ./xformRGBImg cat.ppm cat.rot30c150,150.ppm
translate -150 -150 rotate 30 translate 150 150
$
The output image can be resized with w · √2 × h · √2 to fit any center rotation in.
So, the output image is resized to 425 × 425 where the last translation is adjusted respectively to translate 212.5 212.5:
$ ./xformRGBImg cat.ppm cat.rot30c150,150.425x425.ppm
resize 425 425 translate -150 -150 rotate 30 translate 212.5 212.5
$
The scaling has not yet been checked:
$ ./xformRGBImg cat.ppm cat.rot30c150,150s0.7,0.7.ppm
translate -150 -150 rotate 30 scale 0.7 0.7 translate 150 150
$
Finally, to be fair, I would like to mention the “big brother” of my little toy tool: ImageMagick.
這篇關(guān)于在 C++ 中旋轉(zhuǎn)圖像而不使用 OpenCV 函數(shù)的文章就介紹到這了，希望我們推薦的答案對(duì)大家有所幫助，也希望大家多多支持html5模板網(wǎng)！
【網(wǎng)站聲明】本站部分內(nèi)容來(lái)源于互聯(lián)網(wǎng),旨在幫助大家更快的解決問(wèn)題，如果有圖片或者內(nèi)容侵犯了您的權(quán)益，請(qǐng)聯(lián)系我們刪除處理，感謝您的支持！