| 1 | #include <iostream> |
|---|
| 2 | #include <Utils/utils.hh> |
|---|
| 3 | const int nsample=1000; |
|---|
| 4 | // const int width=300; |
|---|
| 5 | const int width=150; |
|---|
| 6 | const float contrast=0.25; |
|---|
| 7 | |
|---|
| 8 | void zscale(long imagesize, float *image, float &z1, float &z2) |
|---|
| 9 | { |
|---|
| 10 | |
|---|
| 11 | float *smallarray = new float[nsample]; |
|---|
| 12 | float *ct = new float[nsample]; |
|---|
| 13 | long size=0; |
|---|
| 14 | |
|---|
| 15 | float mean=0.,sd=0.; |
|---|
| 16 | for(int i=0;i<imagesize;i++) mean+=image[i]; |
|---|
| 17 | mean /= float(imagesize); |
|---|
| 18 | for(int i=0;i<imagesize;i++) sd+=(image[i]-mean)*(image[i]-mean); |
|---|
| 19 | sd /= float(imagesize); |
|---|
| 20 | |
|---|
| 21 | |
|---|
| 22 | if(imagesize>nsample){ |
|---|
| 23 | int step = (imagesize / nsample) + 1; |
|---|
| 24 | for(int i=0;i<imagesize;i++){ |
|---|
| 25 | if((i%step)==0){ |
|---|
| 26 | smallarray[size] = image[i]; |
|---|
| 27 | ct[size]=(float)size+1.; |
|---|
| 28 | size++; |
|---|
| 29 | } |
|---|
| 30 | } |
|---|
| 31 | } |
|---|
| 32 | else{ |
|---|
| 33 | for(int i=0;i<imagesize;i++){ |
|---|
| 34 | smallarray[i] = image[i]; |
|---|
| 35 | ct[i] = float(i+1); |
|---|
| 36 | } |
|---|
| 37 | size=imagesize; |
|---|
| 38 | } |
|---|
| 39 | |
|---|
| 40 | sort(smallarray,0,size); |
|---|
| 41 | |
|---|
| 42 | /* fit a linear slope to the centre of the cumulative distribution */ |
|---|
| 43 | long midpt = size/2; |
|---|
| 44 | long imin = midpt - width; |
|---|
| 45 | float slope,intercept,errSlope,errIntercept,r; |
|---|
| 46 | if(size<2*width) |
|---|
| 47 | linear_regression(size,ct,smallarray,0,size-1,slope,errSlope, |
|---|
| 48 | intercept,errIntercept,r); |
|---|
| 49 | else linear_regression(size,ct,smallarray,imin,imin+2*width,slope,errSlope, |
|---|
| 50 | intercept,errIntercept,r); |
|---|
| 51 | |
|---|
| 52 | z1 = smallarray[midpt] + (slope/contrast)*(float)(1-midpt); |
|---|
| 53 | z2 = smallarray[midpt] + (slope/contrast)*(float)(nsample-midpt); |
|---|
| 54 | |
|---|
| 55 | if(z1==z2){ |
|---|
| 56 | |
|---|
| 57 | if(z1!=0) z2 = z1 * 1.05; |
|---|
| 58 | else z2 = z1+0.01; |
|---|
| 59 | |
|---|
| 60 | } |
|---|
| 61 | |
|---|
| 62 | delete [] smallarray; |
|---|
| 63 | delete [] ct; |
|---|
| 64 | |
|---|
| 65 | } |
|---|
| 66 | |
|---|
| 67 | |
|---|
| 68 | void zscale(long imagesize, float *image, float &z1, float &z2, float blankVal) |
|---|
| 69 | { |
|---|
| 70 | float *newimage = new float[imagesize]; |
|---|
| 71 | int newsize=0; |
|---|
| 72 | for(int i=0;i<imagesize;i++) if(image[i]!=blankVal) newimage[newsize++] = image[i]; |
|---|
| 73 | |
|---|
| 74 | // cerr<<"Sizes: "<<imagesize<<" "<<newsize<<endl; |
|---|
| 75 | |
|---|
| 76 | float *smallarray = new float[nsample]; |
|---|
| 77 | float *ct = new float[nsample]; |
|---|
| 78 | long size=0; |
|---|
| 79 | |
|---|
| 80 | float mean=0.,sd=0.; |
|---|
| 81 | for(int i=0;i<newsize;i++) mean+=newimage[i]; |
|---|
| 82 | mean /= float(newsize); |
|---|
| 83 | for(int i=0;i<newsize;i++) sd+=(newimage[i]-mean)*(newimage[i]-mean); |
|---|
| 84 | sd /= float(newsize); |
|---|
| 85 | |
|---|
| 86 | |
|---|
| 87 | if(newsize>nsample){ |
|---|
| 88 | int step = (newsize / nsample) + 1; |
|---|
| 89 | for(int i=0;i<newsize;i++){ |
|---|
| 90 | if((i%step)==0){ |
|---|
| 91 | smallarray[size] = newimage[i]; |
|---|
| 92 | ct[size]=(float)size+1.; |
|---|
| 93 | size++; |
|---|
| 94 | } |
|---|
| 95 | } |
|---|
| 96 | } |
|---|
| 97 | else{ |
|---|
| 98 | for(int i=0;i<newsize;i++){ |
|---|
| 99 | smallarray[i] = newimage[i]; |
|---|
| 100 | ct[i] = float(i+1); |
|---|
| 101 | } |
|---|
| 102 | size=newsize; |
|---|
| 103 | } |
|---|
| 104 | |
|---|
| 105 | sort(smallarray,0,size); |
|---|
| 106 | |
|---|
| 107 | /* fit a linear slope to the centre of the cumulative distribution */ |
|---|
| 108 | long midpt = size/2; |
|---|
| 109 | long imin = midpt - width; |
|---|
| 110 | float slope,intercept,errSlope,errIntercept,r; |
|---|
| 111 | if(size<2*width) |
|---|
| 112 | linear_regression(size,ct,smallarray,0,size-1,slope,errSlope, |
|---|
| 113 | intercept,errIntercept,r); |
|---|
| 114 | else linear_regression(size,ct,smallarray,imin,imin+2*width,slope,errSlope, |
|---|
| 115 | intercept,errIntercept,r); |
|---|
| 116 | |
|---|
| 117 | z1 = smallarray[midpt] + (slope/contrast)*(float)(1-midpt); |
|---|
| 118 | z2 = smallarray[midpt] + (slope/contrast)*(float)(nsample-midpt); |
|---|
| 119 | |
|---|
| 120 | if(z1==z2){ |
|---|
| 121 | |
|---|
| 122 | if(z1!=0) z2 = z1 * 1.05; |
|---|
| 123 | else z2 = z1+0.01; |
|---|
| 124 | |
|---|
| 125 | } |
|---|
| 126 | |
|---|
| 127 | delete [] newimage; |
|---|
| 128 | delete [] smallarray; |
|---|
| 129 | delete [] ct; |
|---|
| 130 | |
|---|
| 131 | |
|---|
| 132 | } |
|---|
| 133 | |
|---|
| 134 | |
|---|
