[编程练习] 卷积和实现demo——对阶跃函数进行一维高斯模糊

最新推荐文章于 2024-02-06 21:10:25 发布

原创最新推荐文章于 2024-02-06 21:10:25 发布 · 1.6k 阅读

1 ·

CC 4.0 BY-SA版权

文章标签：

#信号处理 #编程练习 #卷积和

信号处理同时被 3 个专栏收录

1 篇文章

订阅专栏

编程练习

1 篇文章

订阅专栏

卷积和

1 篇文章

订阅专栏

本文介绍了一种使用模板实现曲线生成与处理的技术，包括正态曲线、步进函数曲线、卷积曲线等，通过代码示例展示了如何创建、转换和打印这些曲线。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

#include 
#include 
#include 
#include 
#define _PI_ 3.14159265358979323846264
#define _E_  2.7182818284590452353602874713527

template 
struct Array{
	T * data;
	int length;
	Array(){data=0;length=0;};
	Array(int len){
		printf("allocing %d\n",len);
		data = new T[len];
		length = len;
	};
	Array(Array & a){data = new T[a->length];length = a->length;};
};

template 
struct pair{ T a,b; };

double normal(double x, double u, double sigma){
	return exp((-1*pow(x-u,2))/(2*sigma*sigma))/(sqrtf(2*_PI_)*sigma);
}

Array< pair > * normalCurve(double u, double sigma, int count, float width = 3.0){
	double step = sigma * width * 2.0 /(double)count;
	double ptr = u-sigma*width;
	Array< pair > * curve = new Array< pair >(count);
	for(int i=0;idata[i].a = ptr;
		curve->data[i].b = normal(ptr,u,sigma);
		ptr += step;
	}
	puts("normalCurve");
	return curve;
}

Array * curveTemplate(Array< pair > * a){
	Array * tpl = new Array(a->length);
	for(int i=0;ilength;i++)
		tpl->data[i] = a->data[i].b;
	puts("curveTemplate");
	return tpl;
}

void curvePrt( Array< pair > * curve, FILE * f = stdout){
	for(int i=0;ilength;i++)
		fprintf(f,"%lf,%lf\n",curve->data[i].a,curve->data[i].b);
	puts("curvePrt");
	return;
}

void curvePrt( Array< pair > * curve, const char * filename){
	FILE * f = fopen(filename,"w");
	for(int i=0;ilength;i++)
		fprintf(f,"%lf,%lf\n",curve->data[i].a,curve->data[i].b);
	puts("curvePrt");
	fclose(f);
	return;
}

Array * convolution( Array * a, Array * b ){
	int bstart;
	int bend;
	int blenh = b->length / 2;
	Array * conv = new Array(a->length);
	memset(conv->data,0,a->length*sizeof(double));
	for(int i=0;ilength;i++){
		bstart = (i-blenh)<0 ? (blenh-i) : 0;
		bend   = (i+blenh)>a->length ? (a->length-i+blenh) : b->length;
		for(int j=bstart;jdata[i+j-blenh] += a->data[i]*b->data[j];
		}
	}
	puts("convolution");
	return conv;
}

Array< pair > * makeCurve( Array * a, double start, double step ){
	Array< pair > * curve = new Array< pair >(a->length);
	double prt = start;
	for(int i=0;ilength;i++){
		curve->data[i].a = prt;
		curve->data[i].b = a->data[i];
		prt += step;
	}
	puts("makeCurve");
	return curve;
}

int  main(){
	int stepSize = 2000;
	Array * stepFunc = new Array(stepSize);
	memset(stepFunc->data,0,stepSize*sizeof(double));
	for(int i=stepSize/2-40;idata[i] = 1;
	}
	puts("start");
	curvePrt( normalCurve(0,0.4,100) , "norm.csv" );
	curvePrt( makeCurve( stepFunc,-1*stepSize/2/10.0,1/10.0 ) , "step.csv" );
	curvePrt( makeCurve( convolution( stepFunc,	curveTemplate( normalCurve( 0, 0.4, 100 )) ), -1*stepSize/2/10.0, 1/10.0 ) , "stdout.csv" );
}