Feature Extraction and Object Recognition
Feature extraction and object recognition has been one of the key research issues addressed by researchers and scientists in the fields of computer vision and digital photogrammetry. The creation of automatic map-making machines is the ultimate objective of digital photogrammetry. Research in this area has been focusing on:
- Automatic extraction of road intersections from scanned paper maps (Figure 1).
- Automatic recognition of road signs in color terrestrial imagery. The recognition process is based on the geometric and the radiometric properties of the sought-after objects and also the imaging system (Figure 2).
- Automatic matching and reconstruction of three-dimensional linear features from stereo-imagery. Reconstructed 3-D features can be analyzed for object recognition applications (Figure 3).
Figure 1: Automatic extraction from scanned paper maps.
Figure 2: Road sign recognition in color images.
Figure 3: Automatic matching and 3-D reconstruction of linear features.