Plant Leaves Edge Detection Using Mean-shift Algorithm and Wavelet Transform

Abstract

In this paper, a novel edge detection algorithm based on Mean-shift algorithm and wavelet transform is proposed to accurately segment the Leaves in the Leaf overlapped and grain complex situations. Firstly, Mean-shift algorithm of adaptive spatial bandwidth selection was put forward to smooth the plants leaves image Secondly two-dimensional wavelet transform was utilized to enhance contrast ratio. Then the edge of the plant leaf was lastly detected based on Canny operator. Compared with traditional edge detection algorithms, this edge detected by this method has a better continuity and has a higher signal-to-noise ratio.