//

//  opencv.cpp

//  功能： 图片里取出四边形

//  环境搭建：

//  参考地址：http://vitrum.github.io/2015/07/28/Opencv-%E5%BA%94%E7%94%A8%EF%BC%8C%E5%9B%BE%E7%89%87%E9%87%8C%E5%8F%96%E5%87%BA%E5%9B%9B%E8%BE%B9%E5%BD%A2/

// 

//

#include <opencv2/opencv.hpp>

#include <opencv2/highgui/highgui.hpp>

#include <iostream>

using namespace cv;

using namespace std;

/**

* 边缘检测

* @param gray - grayscale input image

* @param canny - output edge image

*/

void getCanny(Mat gray, Mat &canny) {

    Mat thres;

    double high_thres = threshold(gray, thres, 0, 255, CV_THRESH_BINARY | CV_THRESH_OTSU), low_thres = high_thres * 0.5;

    Canny(gray, canny, low_thres, high_thres);

}

struct Line {

    Point _p1;

    Point _p2;

    Point _center;

    Line(Point p1, Point p2) {

        _p1 = p1;

        _p2 = p2;

        _center = Point((p1.x + p2.x) / 2, (p1.y + p2.y) / 2);

    }

};

bool cmp_y(const Line &p1, const Line &p2) {

    return p1._center.y < p2._center.y;

}

bool cmp_x(const Line &p1, const Line &p2) {

    return p1._center.x < p2._center.x;

}

/**

* Compute intersect point of two lines l1 and l2

* @param l1

* @param l2

* @return Intersect Point

*/

Point2f computeIntersect(Line l1, Line l2) {

    int x1 = l1._p1.x, x2 = l1._p2.x, y1 = l1._p1.y, y2 = l1._p2.y;

    int x3 = l2._p1.x, x4 = l2._p2.x, y3 = l2._p1.y, y4 = l2._p2.y;

    if (float d = (x1 - x2) * (y3 - y4) - (y1 - y2) * (x3 - x4)) {

        Point2f pt;

        pt.x = ((x1 * y2 - y1 * x2) * (x3 - x4) - (x1 - x2) * (x3 * y4 - y3 * x4)) / d;

        pt.y = ((x1 * y2 - y1 * x2) * (y3 - y4) - (y1 - y2) * (x3 * y4 - y3 * x4)) / d;

        return pt;

    }

    return Point2f(-1, -1);

}

void scan(String file, bool debug = false) {

    // 读入原图

    Mat img = imread(file);

    Mat img_proc;  

    int w = img.size().width, h = img.size().height, min_w = 200;

    // 计算缩放比例

    double scale = min(10.0, w * 1.0 / min_w);

    int w_proc = w * 1.0 / scale, h_proc = h * 1.0 / scale;

    // 缩小图片分辨率，提高计算速度

    resize(img, img_proc, Size(w_proc, h_proc));

    Mat img_dis = img_proc.clone();

    /*获取纸的四个边*/

    Mat gray, canny;

    cvtColor(img_proc, gray, CV_BGR2GRAY);

    // 用canny算子进行边缘检测

    getCanny(gray, canny);

    // w_proc 就是缩小后的画面宽度，20是把间距20以内的线段延长拼接为一条直线

    vector<Vec4i> lines;

    v

    HoughLinesP(canny, lines, 1, CV_PI / 180, w_proc / 3, w_proc / 3, 20);

    for (size_t i = 0; i < lines.size(); i++) {

        Vec4i v = lines[i];

        double delta_x = v[0] - v[2], delta_y = v[1] - v[3];

        Line l(Point(v[0], v[1]), Point(v[2], v[3]));

        // get horizontal lines and vertical lines respectively

        if (fabs(delta_x) > fabs(delta_y)) {

            horizontals.push_back(l);

        }

        else {

            verticals.push_back(l);

        }

        // for visualization only

        if (debug)

            line(img_proc, Point(v[0], v[1]), Point(v[2], v[3]), Scalar(0, 0, 255), 1, CV_AA);

    }

    // 边缘情况下，当没有足够的线检测

    if (horizontals.size() < 2) {

        if (horizontals.size() == 0 || horizontals[0]._center.y > h_proc / 2) {

            horizontals.push_back(Line(Point(0, 0), Point(w_proc - 1, 0)));

        }

        if (horizontals.size() == 0 || horizontals[0]._center.y <= h_proc / 2) {

            horizontals.push_back(Line(Point(0, h_proc - 1), Point(w_proc - 1, h_proc - 1)));

        }

    }

    if (verticals.size() < 2) {

        if (verticals.size() == 0 || verticals[0]._center.x > w_proc / 2) {

            verticals.push_back(Line(Point(0, 0), Point(0, h_proc - 1)));

        }

        if (verticals.size() == 0 || verticals[0]._center.x <= w_proc / 2) {

            verticals.push_back(Line(Point(w_proc - 1, 0), Point(w_proc - 1, h_proc - 1)));

        }

    }

    // 按中心点排序

    sort(horizontals.begin(), horizontals.end(), cmp_y);

    sort(verticals.begin(), verticals.end(), cmp_x);

    // for visualization only

    if (debug) {

        line(img_proc, horizontals[0]._p1, horizontals[0]._p2, Scalar(0, 255, 0), 2, CV_AA);

        line(img_proc, horizontals[horizontals.size() - 1]._p1, horizontals[horizontals.size() - 1]._p2, Scalar(0, 255, 0), 2, CV_AA);

        line(img_proc, verticals[0]._p1, verticals[0]._p2, Scalar(255, 0, 0), 2, CV_AA);

        line(img_proc, verticals[verticals.size() - 1]._p1, verticals[verticals.size() - 1]._p2, Scalar(255, 0, 0), 2, CV_AA);

    }

    /* 透视变换 */

    // define the destination image size: A4 - 200 PPI

    int w_a4 = 1654, h_a4 = 2339;

    //int w_a4 = 595, h_a4 = 842;

    Mat dst = Mat::zeros(h_a4, w_a4, CV_8UC3);

    // 求四顶点坐标 corners of destination image with the sequence [tl, tr, bl, br]

    vector<Point2f> dst_pts, img_pts;

    dst_pts.push_back(Point(0, 0));

    dst_pts.push_back(Point(w_a4 - 1, 0));

    dst_pts.push_back(Point(0, h_a4 - 1));

    dst_pts.push_back(Point(w_a4 - 1, h_a4 - 1));

    // corners of source image with the sequence [tl, tr, bl, br]

    img_pts.push_back(computeIntersect(horizontals[0], verticals[0]));

    img_pts.push_back(computeIntersect(horizontals[0], verticals[verticals.size() - 1]));

    img_pts.push_back(computeIntersect(horizontals[horizontals.size() - 1], verticals[0]));

    img_pts.push_back(computeIntersect(horizontals[horizontals.size() - 1], verticals[verticals.size() - 1]));

    // 转换成原始图像比例尺

    for (size_t i = 0; i < img_pts.size(); i++) {

        // for visualization only

        if (debug) {

            circle(img_proc, img_pts[i], 10, Scalar(255, 255, 0), 3);

        }

        img_pts[i].x *= scale;

        img_pts[i].y *= scale;

    }

    // 得到的变换矩阵 用getPerspectiveTransform计算转化矩阵，再用warpPerspective调用转化矩阵进行拉伸

    Mat transmtx = getPerspectiveTransform(img_pts, dst_pts);

    // 应用透视变换

    warpPerspective(img, dst, transmtx, dst.size());

    // 保存照片到本地

    //imwrite("dst.jpg", dst);

    imshow("dst.jpg",dst);

    waitKey(0);

    // for visualization only

    if (debug) {

        namedWindow("dst", CV_WINDOW_KEEPRATIO);

        imshow("src", img_dis);

        imshow("canny", canny);

        imshow("img_proc", img_proc);

        imshow("dst", dst);

        waitKey(0);

    }

}

int main(int argc, char** argv) {

    string img_path[] = { "ceshi.jpg" };

    scan(img_path[0]);

    return 0;

}