数据压缩作业——计算BGR的熵

概述

图像熵：图像熵是一种特征的统计形式，它反映了图像中平均信息量的多少。
图像熵越大，所蕴含的信息越多。
同理，BGR中信息熵大的，所蕴含的信息更多。

---

---

一、公式

二、思路

具体实现思路如下：

1.统计B G R中0~255值出现的频次

代码如下：

    for (int h = 0; h < 256; h++)
	{
		int i = 0;
		for (int w = 0; w < 256 * 3; w = w + 3)
		{
			*(B_buffer + h * 256 + i) = *(rgb + h * 256 * 3 + w);
			*(G_buffer + h * 256 + i) = *(rgb + h * 256 * 3 + w + 1);
			*(R_buffer + h * 256 + i) = *(rgb + h * 256 * 3 + w + 2);
			i++;
		}
	}//将图像中的rgb值分别存放在 * B_buffer, * G_buffer, * R_buffer中；

	for (int g = 0; g < 256; g++)
	{
		for (int j = 0; j < 256 * 256; j++)
		{
			if (*(B_buffer+j) == g)
			{
				B_count[g] ++;
			}
			if (*(G_buffer + j) == g)
			{
				G_count[g] ++;
			}
			if (*(R_buffer + j) == g)
			{
				R_count[g] ++;
			}
		}
	}//遍历3个指针，分别统计B,G,R中0~255这些值出现的次数,其遍历的次数应该是：
	//B,G,R的大小256*256基础上*256就是[0,255]的值
	for (int t = 0; t < 256; t++)
	{
		B_count[t] = B_count[t] / (256 * 256);
		G_count[t] = G_count[t] / (256 * 256);
		R_count[t] = R_count[t] / (256 * 256);
	}//统计B,G,R中0~255这些值的频率

2.利用公式计算BGR的熵

代码如下：

for (int n = 0; n < 256; n++)
	{
		if ( *(B_count + n) != 0)
		{
			B_H[n] = -B_count[n] * (log(B_count[n]) / log(2.0));
		}
		if ( *(G_count + n) != 0)
		{
			G_H[n] = -G_count[n] * (log(G_count[n]) / log(2.0));
		}
		if ( *(R_count + n) != 0)
		{
			R_H[n] = -R_count[n] * (log(R_count[n]) / log(2.0));
		}
	}//计算单个值的熵



	for (int sum = 0; sum < 256; sum++)
	{
		BH += B_H[sum];
		GH += G_H[sum];
		RH += R_H[sum];
	}//计算B,G,R的总熵；

---

总结

B的熵为6.85686
G的熵为7.17846
R的熵为7.22955

完整代码如下：

#include<iostream>
#include<stdio.h>
#include<math.h>
using namespace std;
int main()
{
	FILE* test;
	test = fopen("F:\\大三下资料\\数据压缩\\test.rgb", "rb");
	unsigned char* rgb, * B_buffer, * G_buffer, * R_buffer;
	rgb = (unsigned char*)malloc(256 * 256 * 3 * sizeof(unsigned char));
	B_buffer = (unsigned char*)malloc(256 * 256 * sizeof(unsigned char));
	G_buffer = (unsigned char*)malloc(256 * 256 * sizeof(unsigned char));
	R_buffer = (unsigned char*)malloc(256 * 256 * sizeof(unsigned char));
	double B_count[256] = { 0 };
	double G_count[256] = { 0 };
	double R_count[256] = { 0 };
	double B_H[256] = { 0 };
	double G_H[256] = { 0 };
	double R_H[256] = { 0 };
	double BH = 0;
	double GH = 0;
	double RH = 0;

	fread(rgb, 1, 256 * 256 * 3, test);

	for (int h = 0; h < 256; h++)
	{
		int i = 0;
		for (int w = 0; w < 256 * 3; w = w + 3)
		{
			*(B_buffer + h * 256 + i) = *(rgb + h * 256 * 3 + w);
			*(G_buffer + h * 256 + i) = *(rgb + h * 256 * 3 + w + 1);
			*(R_buffer + h * 256 + i) = *(rgb + h * 256 * 3 + w + 2);
			i++;
		}
	}//将图像中的rgb值分别存放在 * B_buffer, * G_buffer, * R_buffer中；

	for (int g = 0; g < 256; g++)
	{
		for (int j = 0; j < 256 * 256; j++)
		{
			if (*(B_buffer+j) == g)
			{
				B_count[g] ++;
			}
			if (*(G_buffer + j) == g)
			{
				G_count[g] ++;
			}
			if (*(R_buffer + j) == g)
			{
				R_count[g] ++;
			}
		}
	}//遍历3个指针，分别统计B,G,R中0~255这些值出现的次数,其遍历的次数应该是：
	//B,G,R的大小256*256基础上*256就是[0,255]的值

	for (int t = 0; t < 256; t++)
	{
		B_count[t] = B_count[t] / (256 * 256);
		G_count[t] = G_count[t] / (256 * 256);
		R_count[t] = R_count[t] / (256 * 256);
	}//统计B,G,R中0~255这些值的频率


	for (int n = 0; n < 256; n++)
	{
		if ( *(B_count + n) != 0)
		{
			B_H[n] = -B_count[n] * (log(B_count[n]) / log(2.0));
		}
		if ( *(G_count + n) != 0)
		{
			G_H[n] = -G_count[n] * (log(G_count[n]) / log(2.0));
		}
		if ( *(R_count + n) != 0)
		{
			R_H[n] = -R_count[n] * (log(R_count[n]) / log(2.0));
		}
	}//计算单个值的熵



	for (int sum = 0; sum < 256; sum++)
	{
		BH += B_H[sum];
		GH += G_H[sum];
		RH += R_H[sum];
	}//计算B,G,R的总熵；

	cout << "B的熵为" << BH << endl;
	cout << "G的熵为" << GH << endl;
	cout << "R的熵为" << RH << endl;
}