Edge Detection of Plant Root Image Based on Wavelet Transform

摘要

To obtain plant root growing information characters, making edge detection as applicable background, the relevant technology is studied under wavelet transform. Plant root real images are acquired by transparent tubes and image sensors which are bedded in the soil. The methods of selecting the wavelet mother function are analyzed in edge detection, and according to the analytic results the second order B-spline biorthogonal wavelet is selected. This wavelet possesses both compact support and symmetry (or anti-symmetry). By using this method the multiresolution-edge characterizations are extracted. The results of simulation and all practical applications demonstrate the main information of the image is deserved and that the fine trivial edge is prominent by using this method.