Combining Blur and Affine Moment Invariants in Object Recognition

YC Li,HX Chen,JJ Zhang,PF Qu
DOI: https://doi.org/10.1117/12.521525
2003-01-01
Abstract:In many application areas such as military photointerpretation and remote sensing, images are usually subjected to geometric distortion and blur degradation. The determination of invariant characteristics is an important problem in pattern recognition. In this paper, an approach to derive blur and affine combined invariants is presented to recognize the objects. As we prove in the paper, they can be constructed by combining affine moment invariants and blur invariants derived earlier. The tests show that combined invariants can recognize objects in the degraded image without any restoration and geometric normalization.
What problem does this paper attempt to address?