【学习日记】OpenCV-C++ 计算图像的熵值

计算图像熵值的函数：

cv::Scalar Entropy(cv::Mat image)
{
	std::vector<cv::Mat> channels;
	cv::split(image, channels);
	int histSize = 256;
	float range[] = { 0, 256 };
	const float* histRange = { range };
	bool uniform = true;
	bool accumulate = false;
	cv::Mat hist0, hist1, hist2;
	cv::calcHist(&channels[0], 1, 0, cv::Mat(), hist0, 1, &histSize, &histRange, uniform, accumulate);
	cv::calcHist(&channels[1], 1, 0, cv::Mat(), hist1, 1, &histSize, &histRange, uniform, accumulate);
	cv::calcHist(&channels[2], 1, 0, cv::Mat(), hist2, 1, &histSize, &histRange, uniform, accumulate);

	//frequency
	float f0 = 0, f1 = 0, f2 = 0;
	for (int i = 0; i < histSize; i++){
		f0 += hist0.at<float>(i);
		f1 += hist1.at<float>(i);
		f2 += hist2.at<float>(i);	
	}

	//entropy
	cv::Scalar e;
	e.val[0] = 0;
	e.val[1] = 0;
	e.val[2] = 0;
	// e0=0, e1=0, e2=0;
	float p0, p1, p2;

	for (int i = 0; i < histSize; i++){
		p0 = abs(hist0.at<float>(i)) / f0;
		p1 = abs(hist1.at<float>(i)) / f1;
		p2 = abs(hist2.at<float>(i)) / f2;
		if (p0 != 0)
			e.val[0] += -p0 * log10(p0);
		if (p1 != 0)
			e.val[1] += -p1 * log10(p1);
		if (p2 != 0)
			e.val[2] += -p2 * log10(p2);
	}
	return e;
}

在使用时候：用以下方法调用即可

cv::Scalar ent = Entropy(frame);
cout << "-entropy:" << ent.val[0] << endl;