Performance
HALCON is implemented for highest performance, e.g., by actively exploiting multi-core computers, as well as MMX and SSE2. The following runtime examples were measured using a byte image of size 640 x 480 on an Intel Core 2 / 3 GHz computer. Note: runtime may vary with different input data.
| Shape-based matching with a template of size 100 x 100 (the search area is the complete image with 360° rotation) |
3.2 ms |
| Affine transformation (nearest neighbor) | 0.6 ms |
| Sobel edge filter | 0.6 ms |
| Median (3 x 3) | 0.4 ms |
| Binomial filter (5 x 5) | 0.8 ms |
| Gray opening (3 x 3) | 0.3 ms |
| Binary dilation (50 x 50) | 0.1 ms |
| Binary erosion (50 x 50) | 0.04 ms |
| Threshold operation | 0.1 ms |
| Subpixel-accurate threshold | 1.1 ms |
| Feature calculation for 250 objects (blobs) (features: 'center of gravity' & 'number of pixels') |
0.04 ms |
| Subpixel-accurate measuring of edge positions (search size 100 x 20) |
0.01 ms |
| Fast Fourier transform | 4.3 ms |
