Sub-pixel detection algorithm based on cubic B-spline curve and multi-scale adaptive wavelet transform

This paper presents a method of sub-pixel image edge detection based on the cubic B-spline curve interpolation and the wavelet transform, to solve the problem that the system has increasing demands to the algorithm in real-time and the accuracy. The high resolution and anti-noise of the multi-scale adaptive threshold wavelet transform are used to test the modulus maxima of the image edge. Using the cubic B-spline curve interpolation, the accurate interpolation and the edge of the image reconstruction can be realized in sub-pixel level when the image edge is irregular. The experimental results indicate that mean squared error in the new method is less than 0.02 pixel. The measure data of the standard deviation is 0.0043 mm. From above on, the proposed method is robust to noise and it has a high-precision positioning. Compared to these popular edge detection algorithms, the proposed method performs better than those algorithms on detail reserving and positioning accuracy.