OpenCV学习笔记之扫描图像

亦凉 2023-07-14 10:44 1阅读 0赞

# 一、 图像在内存中存储方式 #

图像矩阵的大小取决于我们所用的颜色模型，确切地说，取决于所用通道数。如果是灰度图像，矩阵就会像这样：

![\\newcommand\{\\tabItG\}\[1\] \{ \\textcolor\{black\}\{\#1\} \\cellcolor\[gray\]\{0.8\}\} \\begin\{tabular\} \{ccccc\} ~ & \\multicolumn\{1\}\{c\}\{Column 0\} &   \\multicolumn\{1\}\{c\}\{Column 1\} &   \\multicolumn\{1\}\{c\}\{Column ...\} & \\multicolumn\{1\}\{c\}\{Column m\}\\\\ Row 0 & \\tabItG\{0,0\} & \\tabItG\{0,1\} & \\tabItG\{...\}  & \\tabItG\{0, m\} \\\\ Row 1 & \\tabItG\{1,0\} & \\tabItG\{1,1\} & \\tabItG\{...\}  & \\tabItG\{1, m\} \\\\ Row ... & \\tabItG\{...,0\} & \\tabItG\{...,1\} & \\tabItG\{...\} & \\tabItG\{..., m\} \\\\ Row n & \\tabItG\{n,0\} & \\tabItG\{n,1\} & \\tabItG\{n,...\} & \\tabItG\{n, m\} \\\\ \\end\{tabular\}][newcommand_tabItG_1_ _ _textcolor_black_1_ _cellcolor_gray_0.8_ _begin_tabular_ _ccccc_ _ _ _multicolumn_1_c_Column 0_ _ _multicolumn_1_c_Column 1_ _ _multicolumn_1_c_Column ..._ _ _multicolumn_1_c_Column m_ Row 0 _ _tabItG_0_0_ _ _tabItG_0_1_ _ _tabItG_..._ _ _tabItG_0_ m_ _ Row 1 _ _tabItG_1_0_ _ _tabItG_1_1_ _ _tabItG_..._ _ _tabItG_1_ m_ _ Row ... _ _tabItG_..._0_ _ _tabItG_..._1_ _ _tabItG_..._ _ _tabItG_..._ m_ _ Row n _ _tabItG_n_0_ _ _tabItG_n_1_ _ _tabItG_n_..._ _ _tabItG_n_ m_ _ _end_tabular]

而对多通道图像来说，矩阵中的列会包含多个子列，其子列个数与通道数相等。例如，RGB颜色模型的矩阵：

![\\newcommand\{\\tabIt\}\[1\] \{ \\textcolor\{yellow\}\{\#1\} \\cellcolor\{blue\} &  \\textcolor\{black\}\{\#1\} \\cellcolor\{green\} & \\textcolor\{black\}\{\#1\} \\cellcolor\{red\}\} \\begin\{tabular\} \{ccccccccccccc\} ~ & \\multicolumn\{3\}\{c\}\{Column 0\} &   \\multicolumn\{3\}\{c\}\{Column 1\} &   \\multicolumn\{3\}\{c\}\{Column ...\} & \\multicolumn\{3\}\{c\}\{Column m\}\\\\ Row 0 & \\tabIt\{0,0\} & \\tabIt\{0,1\} & \\tabIt\{...\}  & \\tabIt\{0, m\} \\\\ Row 1 & \\tabIt\{1,0\} & \\tabIt\{1,1\} & \\tabIt\{...\}  & \\tabIt\{1, m\} \\\\ Row ... & \\tabIt\{...,0\} & \\tabIt\{...,1\} & \\tabIt\{...\} & \\tabIt\{..., m\} \\\\ Row n & \\tabIt\{n,0\} & \\tabIt\{n,1\} & \\tabIt\{n,...\} & \\tabIt\{n, m\} \\\\ \\end\{tabular\}][newcommand_tabIt_1_ _ _textcolor_yellow_1_ _cellcolor_blue_ _ _textcolor_black_1_ _cellcolor_green_ _ _textcolor_black_1_ _cellcolor_red_ _begin_tabular_ _ccccccccccccc_ _ _ _multicolumn_3_c_Column 0_ _ _multicolumn_3_c_Column 1_ _ _multicolumn_3_c_Column ..._ _ _multicolumn_3_c_Column m_ Row 0 _ _tabIt_0_0_ _ _tabIt_0_1_ _ _tabIt_..._ _ _tabIt_0_ m_ _ Row 1 _ _tabIt_1_0_ _ _tabIt_1_1_ _ _tabIt_..._ _ _tabIt_1_ m_ _ Row ... _ _tabIt_..._0_ _ _tabIt_..._1_ _ _tabIt_..._ _ _tabIt_..._ m_ _ Row n _ _tabIt_n_0_ _ _tabIt_n_1_ _ _tabIt_n_..._ _ _tabIt_n_ m_ _ _end_tabular]

注意到，子列的通道顺序是反过来的：BGR而不是RGB。很多情况下，因为内存足够大，可实现连续存储，因此，图像中的各行就能一行一行地连接起来，形成一个长行。连续存储有助于提升图像扫描速度，我们可以使用 [isContinuous()][isContinuous] 来去判断矩阵是否是连续存储的. 相关示例会在接下来的内容中提供。

# 二、扫描图像几种方式 #

## 2.1 指针访问 ##

/*!
     * \brief scanImagePtr
     * \param image
     * \return
     */
    Mat& scanImagePtr(Mat &image){
        double start = static_cast<double>(getTickCount());
        //! 行
        int rowNumber = image.rows;
        //! 每一行元素个数 = 列数 x 通道数
        int colNumber = image.cols * image.channels();
        for (int i = 0; i < rowNumber; i++)
        {
            //! 获取第i行的首地址
            uchar* data = image.ptr<uchar>(i);
            for (int j = 0; j < colNumber; j++)
            {
                data[j] /= 2;
            }
        }
        double end = static_cast<double>(getTickCount());
        double time = (end - start) / getTickFrequency();
        cout << "type1:" << time << "s" << endl;
        return image;
    }

## 2.2 迭代器iterator ##

/*!
     * \brief scanImageiterator
     * \param image
     * \return
     */
    Mat& scanImageiterator(Mat &image){
        double start = static_cast<double>(getTickCount());
        //! 初始位置的迭代器
        Mat_<Vec3b>::iterator it = image.begin<Vec3b>();
        //! 终止位置的迭代器
        Mat_<Vec3b>::iterator itend = image.end<Vec3b>();
        for (; it != itend; it++)
        {
            //! 处理BGR三个通道
            (*it)[0] = 255;//B
            //(*it)[1] = 255;//G
            //(*it)[2] = 0;//R
        }
        double end = static_cast<double>(getTickCount());
        double time = (end - start) / getTickFrequency();//计算时间
        cout << "type2:" << time << "s" << endl;
        return image;
    }

## 2.3 动态地址计算 ##

/*!
     * \brief scanImageAt
     * \param image
     * \return
     */
    Mat& scanImageAt(Mat &image){
        double start = static_cast<double>(getTickCount());
        int rowNumber = image.rows;
        int colNumber = image.cols;
        for (int i = 0; i < rowNumber; i++)
            for (int j = 0; j < colNumber; j++)
            {
                //! 处理BGR三个通道
                //image.at<Vec3b>(i, j)[0] = 0;//B
                image.at<Vec3b>(i, j)[1] = 255;//G
                //image.at<Vec3b>(i, j)[2] = 0;//R
            }
        double end = static_cast<double>(getTickCount());
        double time = (end - start) / getTickFrequency();//计算时间
        cout << "type3:" << time << "s" << endl;
        return image;
    }

## 2.4 三种方法效能对比 ##

type1:0.000189559s
    type2:0.000832323s
    type3:0.000424337s

经典的C风格运算符\[\]（指针）访问要更胜一筹。

代码详见《[OpenCV扫描图像三种方式练习][OpenCV]》

# 三、OpenCv坐标系 #

[OpenCV][OpenCV 1]中坐标体系中的零点坐标定义为图片的左上角，X轴为图像矩形的上面那条水平线，从左往右；Y轴为图像矩形左边的那条垂直线，从上往下。在Point(x,y)和Rect(x,y)中，第一个参数x代表的是元素所在图像的列数，第二个参数y代表的是元素所在图像的行数，而在at(x,y)中是相反的。

![SouthEast][]

# 参考资料： #

*  [OpenCV坐标系与操作像素的四种方法][OpenCV 2]
 *  [OpenCV如何扫描图像、利用查找表和计时][OpenCV 3]
 *  [Opencv中三种操作像素的方法][Opencv]

[newcommand_tabItG_1_ _ _textcolor_black_1_ _cellcolor_gray_0.8_ _begin_tabular_ _ccccc_ _ _ _multicolumn_1_c_Column 0_ _ _multicolumn_1_c_Column 1_ _ _multicolumn_1_c_Column ..._ _ _multicolumn_1_c_Column m_ Row 0 _ _tabItG_0_0_ _ _tabItG_0_1_ _ _tabItG_..._ _ _tabItG_0_ m_ _ Row 1 _ _tabItG_1_0_ _ _tabItG_1_1_ _ _tabItG_..._ _ _tabItG_1_ m_ _ Row ... _ _tabItG_..._0_ _ _tabItG_..._1_ _ _tabItG_..._ _ _tabItG_..._ m_ _ Row n _ _tabItG_n_0_ _ _tabItG_n_1_ _ _tabItG_n_..._ _ _tabItG_n_ m_ _ _end_tabular]: /images/20230529/320c824faae84d2ab34e34e1b2e42d5b.png
[newcommand_tabIt_1_ _ _textcolor_yellow_1_ _cellcolor_blue_ _ _textcolor_black_1_ _cellcolor_green_ _ _textcolor_black_1_ _cellcolor_red_ _begin_tabular_ _ccccccccccccc_ _ _ _multicolumn_3_c_Column 0_ _ _multicolumn_3_c_Column 1_ _ _multicolumn_3_c_Column ..._ _ _multicolumn_3_c_Column m_ Row 0 _ _tabIt_0_0_ _ _tabIt_0_1_ _ _tabIt_..._ _ _tabIt_0_ m_ _ Row 1 _ _tabIt_1_0_ _ _tabIt_1_1_ _ _tabIt_..._ _ _tabIt_1_ m_ _ Row ... _ _tabIt_..._0_ _ _tabIt_..._1_ _ _tabIt_..._ _ _tabIt_..._ m_ _ Row n _ _tabIt_n_0_ _ _tabIt_n_1_ _ _tabIt_n_..._ _ _tabIt_n_ m_ _ _end_tabular]: /images/20230529/803906a798a84f1c817a2b183ca32b6b.png
[isContinuous]: http://opencv.itseez.com/modules/core/doc/basic_structures.html#mat-iscontinuous
[OpenCV]: https://blog.csdn.net/a8039974/article/details/104861005
[OpenCV 1]: http://lib.csdn.net/base/opencv
[SouthEast]: /images/20230529/7a234d2f44d7456ba12a3a7aa31a30a6.png
[OpenCV 2]: https://blog.csdn.net/dcrmg/article/details/51843295
[OpenCV 3]: http://www.opencv.org.cn/opencvdoc/2.3.2/html/doc/tutorials/core/how_to_scan_images
[Opencv]: https://blog.csdn.net/a8039974/article/details/80914958