Opencv-图像积分图算法

知识点

图像积分图算法
概述
积分图像是Crow在1984年首次提出，是为了在多尺度透视投影中提高渲染速度，是一种快速计算图像区域和与平方和的算法。其核心思想是对每个图像建立自己的积分图查找表，在图像积分处理计算阶段根据预先建立的积分图查找表，直接查找从而实现对均值卷积线性时间计算，做到了卷积执行的时间与半径窗口大小的无关联。图像积分图在图像特征提取HAAR/SURF、二值图像分析、图像相似相关性NCC计算、图像卷积快速计算等方面均有应用，是图像处理中的经典算法之一。

图像积分图建立与查找
在积分图像(Integral Image - ii)上任意位置(x, y)处的ii(x, y)表示该点左上角所有像素之和， 其中(x,y)是图像像素点坐标。


OpenCV中的相关API如下：
integral(
InputArray src, // 输入图像
OutputArray sum, // 和表
OutputArray sqsum, // 平方和表
OutputArray tilted, // 瓦块和表
int sdepth = -1, // 和表数据深度常见CV_32S
int sqdepth = -1 // 平方和表数据深度 常见 CV_32F
)

py代码

import cv2 as cv
import numpy as np


def get_block_sum(ii, x1, y1, x2, y2, index):
    tl = ii[y1, x1][index]
    tr = ii[y2, x1][index]
    bl = ii[y1, x2][index]
    br = ii[y2, x2][index]
    s = (br - bl - tr + tl)
    return s


def blur_demo(image, ii):
    h, w, dims = image.shape
    result = np.zeros(image.shape, image.dtype)
    ksize = 15
    radius = ksize // 2
    for row in range(0, h + radius, 1):
        y2 = h if (row + 1)> h else (row + 1)
        y1 = 0 if (row - ksize) < 0 else (row - ksize)
        for col in range(0, w + radius, 1):
            x2 = w if (col + 1)>w else (col + 1)
            x1 = 0 if (col - ksize) < 0 else (col - ksize)
            cx = 0 if (col - radius) < 0 else (col - radius)
            cy = 0 if (row - radius) < 0 else (row - radius)
            num = (x2 - x1)*(y2 - y1)
            for i in range(0, 3, 1):
                s = get_block_sum(ii, x1, y1, x2, y2, i)
                result[cy, cx][i] = s // num

    cv.imshow("integral fast blur", result)
    # cv.imwrite("D:/result.png", result)
src = cv.imread("C:/Users/qqxd/Desktop/opencvcode/images/test1.png")
cv.namedWindow("input", cv.WINDOW_AUTOSIZE)
cv.imshow("input", src)
sum_table = cv.integral(src, sdepth=cv.CV_32S)
print(sum_table.shape)
blur_demo(src, sum_table)

cv.waitKey(0)
cv.destroyAllWindows()

c++代码

#include <opencv2/opencv.hpp>
#include <iostream>
using namespace cv;
using namespace std;
void blur_demo(Mat &image, Mat &sum);
void edge_demo(Mat &image, Mat &sum);
int getblockSum(Mat &sum, int x1, int y1, int x2, int y2, int i);
int main(int argc, char** argv) {
	Mat src = imread("C:/Users/qqxd/Desktop/opencvcode/images/test1.png");
	if (src.empty()) {
		printf("could not load image...\n");
		return -1;
	}
	namedWindow("input", CV_WINDOW_AUTOSIZE);
	imshow("input", src);
	namedWindow("output", CV_WINDOW_AUTOSIZE);
	// 计算积分图
	Mat sum, sqrsum;
	integral(src, sum, sqrsum, CV_32S, CV_32F);
	// 积分图应用
	int type = 0;
	while (true) {
		char c = waitKey(100);
		if (c > 0) {
			type = (int)c;
			printf("c : %d\n", type);
		}
		if (c == 27) {
			break; // ESC
		}
		if (type == 49) { // 数字键 1
			blur_demo(src, sum);
		}
		else if (type == 50) { // 数字键 2
			edge_demo(src, sum);
		}
		else {
			blur_demo(src, sum);
		}
	}
	waitKey(0);
	return 0;
}
void blur_demo(Mat &image, Mat &sum) {
	int w = image.cols;
	int h = image.rows;
	Mat result = Mat::zeros(image.size(), image.type());
	int x2 = 0, y2 = 0;
	int x1 = 0, y1 = 0;
	int ksize = 5;
	int radius = ksize / 2;
	int ch = image.channels();
	int cx = 0, cy = 0;
	for (int row = 0; row < h + radius; row++) {
		y2 = (row + 1)>h ? h : (row + 1);
		y1 = (row - ksize) < 0 ? 0 : (row - ksize);
		for (int col = 0; col < w + radius; col++) {
			x2 = (col + 1)>w ? w : (col + 1);
			x1 = (col - ksize) < 0 ? 0 : (col - ksize);
			cx = (col - radius) < 0 ? 0 : col - radius;
			cy = (row - radius) < 0 ? 0 : row - radius;
			int num = (x2 - x1)*(y2 - y1);
			for (int i = 0; i < ch; i++) {
				// 积分图查找和表，计算卷积
				int s = getblockSum(sum, x1, y1, x2, y2, i);
				result.at<Vec3b>(cy, cx)[i] = saturate_cast<uchar>(s / num);
			}
		}
	}
	imshow("output", result);
	imwrite("D:/result.png", result);
}
/**
* 3x3 sobel 垂直边缘检测演示
*/
void edge_demo(Mat &image, Mat &sum) {
	int w = image.cols;
	int h = image.rows;
	Mat result = Mat::zeros(image.size(), CV_32SC3);
	int x2 = 0, y2 = 0;
	int x1 = 0, y1 = 0;
	int ksize = 3; // 算子大小，可以修改，越大边缘效应越明显
	int radius = ksize / 2;
	int ch = image.channels();
	int cx = 0, cy = 0;
	for (int row = 0; row < h + radius; row++) {
		y2 = (row + 1)>h ? h : (row + 1);
		y1 = (row - ksize) < 0 ? 0 : (row - ksize);
		for (int col = 0; col < w + radius; col++) {
			x2 = (col + 1)>w ? w : (col + 1);
			x1 = (col - ksize) < 0 ? 0 : (col - ksize);
			cx = (col - radius) < 0 ? 0 : col - radius;
			cy = (row - radius) < 0 ? 0 : row - radius;
			int num = (x2 - x1)*(y2 - y1);
			for (int i = 0; i < ch; i++) {
				// 积分图查找和表，计算卷积
				int s1 = getblockSum(sum, x1, y1, cx, y2, i);
				int s2 = getblockSum(sum, cx, y1, x2, y2, i);
				result.at<Vec3i>(cy, cx)[i] = saturate_cast<int>(s2 - s1);
			}
		}
	}
	Mat dst, gray;
	convertScaleAbs(result, dst);
	normalize(dst, dst, 0, 255, NORM_MINMAX);
	cvtColor(dst, gray, COLOR_BGR2GRAY);
	imshow("output", gray);
	imwrite("D:/edge_result.png", gray);
}
int getblockSum(Mat &sum, int x1, int y1, int x2, int y2, int i) {
	int tl = sum.at<Vec3i>(y1, x1)[i];
	int tr = sum.at<Vec3i>(y2, x1)[i];
	int bl = sum.at<Vec3i>(y1, x2)[i];
	int br = sum.at<Vec3i>(y2, x2)[i];
	int s = (br - bl - tr + tl);
	return s;
}

运行结果如下: