Boundary identification for wood defects using Neural Network based on Matlab programming

Here we present a novel method to detect the boundaries of wood defects in X-ray wood images using the Hopfield Neural Network (HNN) algorithm based on Matlab programming. An improved energy function was designed for the HNN to perform the boundary detection which was formulated, in this paper, as an optimization process that could locate the boundary points of wood defects. The gray value of each pixel in the image was considered as a neuron state of HNN. An initial defect boundary was roughly estimated using Canny Algorithm. Based on the initial states, all the neuron´s states updated till the energy function get the minimum value. The defect boundary was identified when the neurons get the final states. The results show we obtain a more accurate boundary using this method comparing to using other traditional methods, and the noises were effectively removed. In this paper, we also find a way to do the boundary detection based on Matlab programming.