Image reconstruction
10 views (last 30 days)
Show older comments
Hi all,
I am new with Matlab, I have this 20 x 36 data that represent a transverse image from a phantom. I have it run through iradon code but the image is very unsatisfactory. The code is as the following:
sumdeg4(1:36,1:20)=sum(fourdeg,2); sumdeg4g=mat2gray(sumdeg4'); % arrange projections as sinogram. figure, imshow(sumdeg4g) I=iradon(sumdeg4g,0:10:350) I=mat2gray(I); figure, imshow(I)
The 36 set of data is actually collected from 36 angular projection of 10 degree. Each projection have 20 data on it which is collected from a strip of 20 detector sized 5 mm. I expect the image to be a circle with one hotspot on it. Nonetheless am unable to produce the image. I have tried to use filters to enhance the image but the result is the same. Since I am not really sure how to make sure my work whether it is correct or not and what is the correct way of handling the data, I really hope if there is any of the Matlab expert can shed a light on this problem. The data is:
68 51 83 120 189 218 346 550 1103 4041 4113 1194 555 330 202 165 135 83 69 34
67 90 99 164 205 311 586 1020 3280 5718 1363 565 357 247 144 140 91 57 52 36
67 94 147 206 256 488 909 2370 615 1488 625 392 224 168 122 98 99 57 48 35
94 137 164 265 459 751 1705 567 1616 725 414 257 183 151 103 82 75 57 40 27
113 178 231 331 585 1104 3301 2433 723 416 266 216 112 99 78 64 43 27 36 37
146 187 285 432 756 1981 3677 961 510 282 208 144 92 86 61 56 61 46 19 29
164 223 289 556 970 381 1459 583 346 225 141 103 77 84 55 48 54 21 23 16
185 213 331 566 1380 1766 764 440 260 165 137 88 56 57 63 39 34 35 22 13
182 257 366 583 1302 1836 520 325 186 134 127 95 69 56 39 29 43 41 15 12
146 214 336 518 1264 1193 430 259 199 144 103 93 55 54 43 26 26 34 22 29
134 189 303 419 848 1240 378 255 170 105 83 82 73 47 41 39 34 26 15 25
129 166 235 323 653 1380 448 263 147 130 86 76 74 60 46 38 28 19 18 20
92 153 187 247 418 1263 480 247 165 103 82 100 48 56 38 29 22 35 21 15
81 92 150 197 256 470 852 262 185 141 80 66 75 58 51 30 36 29 25 21
57 91 115 151 193 263 813 374 192 122 101 96 77 45 43 48 47 34 26 17
71 65 57 122 136 176 350 1830 326 167 133 82 69 68 55 35 38 26 33 26
42 45 78 88 94 120 165 327 755 237 146 110 90 75 46 53 44 35 30 18
28 33 63 68 81 135 132 189 368 547 217 122 104 74 69 49 44 43 26 17
36 46 32 50 59 84 109 118 183 357 420 202 134 105 90 75 50 41 23 30
30 33 36 43 58 55 73 110 114 216 541 321 201 126 99 79 72 67 38 42
18 16 31 35 37 54 70 78 128 161 219 815 310 182 124 111 97 75 51 47
16 24 34 28 45 69 56 72 89 119 168 341 1836 377 180 135 114 64 63 66
14 27 22 32 28 49 58 64 78 113 144 180 385 871 272 175 117 103 85 44
15 30 24 28 44 43 55 63 65 85 122 171 297 865 469 273 188 134 128 69
17 28 23 28 34 39 48 68 64 116 137 158 270 498 1253 413 251 183 138 91
16 16 21 32 28 36 57 74 64 98 131 148 238 424 1461 631 301 244 181 121
20 19 18 37 31 34 51 55 55 85 145 198 256 414 1203 910 420 287 203 131
14 16 18 32 46 37 45 69 66 87 118 183 275 474 1210 1216 503 304 233 188
10 24 26 39 52 42 51 75 103 107 159 231 325 581 1738 1267 613 346 234 175
14 19 20 33 37 48 49 74 117 126 183 247 398 756 1674 1407 560 342 224 145
32 34 28 32 48 50 90 105 127 157 199 330 561 1445 401 1036 543 314 204 146
30 36 32 44 53 57 94 111 188 196 301 529 923 3765 2103 761 450 309 212 135
27 34 44 56 59 88 98 119 181 285 406 773 2421 3362 1165 587 390 258 144 112
23 47 74 62 65 96 121 164 262 439 686 1678 561 1631 699 453 306 188 146 105
36 51 59 65 96 99 188 240 361 703 1429 619 2339 942 474 268 183 128 91 81
36 60 80 100 120 156 251 359 588 1283 5699 3335 1046 512 346 245 146 107 93 67
Regards, pejo
0 Comments
Answers (1)
Bjorn Gustavsson
on 26 Mar 2012
My quick guess is that you have too sparse data. The resolution of the inverse radon transform depends on the number of projections you have. 36 projections with 10 degrees step really reduces to 18! (radon(V,phi,l)==radon(V,phi+pi,-l) free-notationish). Further with only 20 line-integrals per each projection you have very little data. If I'm right you could make a simple check by looking at the examples given in the help to radon/iradon, and reduce the number of projections and their resolution to match your case, then run the example with that reduced data. This will give you a sense for what resolution you can expect.
HTH
1 Comment
Ayush singhal
on 20 Apr 2021
how to reconstruct an image if we have pixel points.
is there any code for this or any technique?
( these pixel points are derived from a striping pattern on 3D object).
ANy lead would be helpful.
See Also
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!