faq:technical:what_is_the_algorithm_used_in_find_edges

This shows you the differences between two versions of the page.

— |
faq:technical:what_is_the_algorithm_used_in_find_edges [2019/04/12 13:13] (current) |
||
---|---|---|---|

Line 1: | Line 1: | ||

+ | ====== FAQ: What is the algorithm used in Find Edges? ====== | ||

+ | It is the 3x3 [[http://homepages.inf.ed.ac.uk/rbf/HIPR2/sobel.htm|Sobel edge filter]]. | ||

+ | |||

+ | The source is in the //filter()// method of the// ij/process/ByteProcessor.java //class. | ||

+ | |||

+ | ... | ||

+ | case FIND_EDGES: // 3x3 Sobel filter | ||

+ | sum1 = p1 + 2*p2 + p3 - p7 - 2*p8 - p9; | ||

+ | sum2 = p1 + 2*p4 + p7 - p3 - 2*p6 - p9; | ||

+ | sum = (int)Math.sqrt(sum1*sum1 + sum2*sum2); | ||

+ | if (sum> 255) sum = 255; | ||

+ | break; | ||

+ | ... | ||

+ | |||

+ | p1...p9 hold the grey value of the pixels in the 3x3 kernel (the central pixel is p5). | ||

+ | |||

+ | There are many other edge detecting algorithms (such as Canny, Roberts, Marr's Laplacian-of-Gaussian or LoG, difference-of-boxes or DoB, Frei-Chen, etc.). |

faq/technical/what_is_the_algorithm_used_in_find_edges.txt ยท Last modified: 2019/04/12 13:13 (external edit)