The Corner Detection Algorithm Based on 2-D Discrete Wavelet Transform

Corner detection is a critical and basic task in computer vision and image processing. By far the existing boundary-based wavelet transform approaches can not directly be utilized in gradient image because the theories are mainly about 1-D wavelet transform. We put forward a novel corner detection technique based on 2-D discrete wavelet transform. The corners are extracted by several preset thresholds. Experimental results indicate that this algorithm is not only easy computation and efficient location but is immune to noise. Comparison with a recently proposed technique is also provided.