Minimum distance between the contour pixels of two regions each.
The operator distance_rr_min calculates the minimum distance of pairs of regions. If several regions are passed in Regions1 and Regions2 the distance between the contour pixels of each i-th element is calculated and then forms the i-th entry in the output parameter MinDistance. The Euclidean distance is used. The parameters (Row1, Column1) and (Row2, Column2) indicate the position on the contour of Regions1 and Regions2, respectively, that have the minimum distance.
The calculation is carried out by comparing all contour pixels (see get_region_contour). This means in particular that each region must consist of exactly one connected component and that holes in the regions are ignored. Furthermore, it is not checked whether one region lies completely within the other region. In this case, a minimum distance > 0 is returned. It is also not checked whether both regions contain a nonempty intersection. In the latter case, a minimum distance of 0 or > 0 can be returned, depending on whether the contours of the regions contain a common point or not.
Both input parameters must contain the same number of regions. The regions must not be empty.
|
Regions1 (input_object) |
region(-array) -> object |
| Regions to be examined. | |
|
Regions2 (input_object) |
region(-array) -> object |
| Regions to be examined. | |
|
MinDistance (output_control) |
real(-array) -> real |
| Minimum distance between contours of the regions. | |
| Assertion: 0 <= MinDistance | |
|
Row1 (output_control) |
point.y(-array) -> integer |
| Line index on contour in Regions1. | |
|
Column1 (output_control) |
point.x(-array) -> integer |
| Column index on contour in Regions1. | |
|
Row2 (output_control) |
point.y(-array) -> integer |
| Line index on contour in Regions2. | |
|
Column2 (output_control) |
point.x(-array) -> integer |
| Column index on contour in Regions2. | |
If N1,N2 are the lengths of the contours the runtime complexity is O(N1 * N2).
The operator distance_rr_min returns the value 2 (H_MSG_TRUE) if the input is not empty. Otherwise an exception handling is raised.
distance_rr_min is reentrant and automatically parallelized (on tuple level).
threshold, regiongrowing, connection
distance_rr_min_dil, dilation1, intersection
Foundation
