基于霍夫变换的航迹起始算法研究（Matlab代码实现）

%标准Hough变换
%Author:Shen Baoyin
%Time:2018/8/1
close all
clear all
​
target=2;%目标数
n=15;%起始拍数
k=90;%sig分的个数
m=500;%p分的个数
Monte_Carlo=100;%Monte_Carlo仿真次数
L=150;%雷达量测距离
Pd=1;%检测概率
​
%目标起始坐标及速度
x1=40;y1=20;vx1=0.3;vy1=0.18;%单位km,km/s
x2=20;y2=80;vx2=0.3;vy2=-0.18;
Ts=4;%采样周期，单位s
​
success=zeros(Monte_Carlo,target);%目标航迹成功起始矩阵
fake(1:Monte_Carlo)=0;%目标航迹虚假起始矩阵
track_number(1:Monte_Carlo)=0;%总航迹起始数
​
N=0:n-1;
X1_init=x1+Ts*N*vx1;%真实航迹1
Y1_init=y1+Ts*N*vy1;
Y1_0=y1-vy1*x1/vx1;
offset(1)=Y1_0*cos(atan(abs(vy1)/abs(vx1)));%航迹1真实垂距
X2_init=x2+Ts*N*vx2;%真实航迹2
Y2_init=y2+Ts*N*vy2;
Y2_0=y2-vy2*x2/vx2;
offset(2)=Y2_0*cos(atan(abs(vy2)/abs(vx2)));%航迹2真实垂距
​
Np=1:k;
dNp=pi/k;%参数空间角度间隔
angle=(Np-1/2)*dNp;
​
dMp=6*0.1;%%参数空间垂距间隔
​
for monte=1:Monte_Carlo
    clear R Rn A0 P0 R X_za Y_za noisex noisey
    R = poissrnd(50,1,n);%每拍杂波个数，服从泊松分布
    Rn=R(1);
    X_za=unifrnd (0, 100, 1, R(1));%随机产生（x,y）坐标，服从0-100的均匀分布
    Y_za=unifrnd (0, 100, 1, R(1));
    for i=2:n
        X_za(Rn+1:Rn+R(i))=unifrnd (0, 100, 1, R(i));
        Y_za(Rn+1:Rn+R(i))=unifrnd (0, 100, 1, R(i));
        Rn=Rn+R(i);
    end
​
    noisex=normrnd(0,0.1,1,n);%x量测噪声
    noisey=normrnd(0,0.1,1,n);
​
    X1=X1_init+noisex;X2=X2_init+noisex;%实际量测
    Y1=Y1_init+noisey;Y2=Y2_init+noisey;
​
    A=zeros(k,2*m);%积累矩阵
​
    %航迹1 Hough变换
    for i=1:n
        for j=1:k
            P(i,j)=X1(i)*cos(angle(j))+Y1(i)*sin(angle(j));
        end
    end
    %航迹2 Hough变换
    for i=(n+1):(2*n)
        for j=1:k
            P(i,j)=X2(i-n)*cos(angle(j))+Y2(i-n)*sin(angle(j));
        end
    end
​
    %杂波的Hough变换
    for i=2*n+1:(2*n+Rn)
        for j=1:k
            P(i,j)=X_za(i-2*n)*cos(angle(j))+Y_za(i-2*n)*sin(angle(j));
        end
    end
​
    %对积累矩阵投票
    for i=1:k
        for j=1:2*m
            a=-L+(j-1)*dMp;
            b=-L+j*dMp;
           for h=1:2*n+Rn
               if (P(h,i)>=a && P(h,i)<b) 
                   A(i,j)=A(i,j)+1;
               end
           end
        end
    end