Abstract:
Imagebased intelligent wood identification method is expected to identify wood by automatically extracting identification features of the wood from its images, and is very important to wood sciences and industries. We proposed a morphological feature extraction method of wood pores based on an improved growing region algorithm. With this method, we can realize precise segmentation of wood pore cells from micrographs and acquire ten morphological features of pore cells according to the technology of chain codes tracking. We validated this method in six kinds of micrograph of broadleaf wood. The simulation experiment shows that this algorithm could improve the computational speed of segmentation of wood pores. In the meanwhile, the ten morphological features of pore cells have a quite distinguishable capacity in the six kinds of broadleaf wood. It is suggested that the algorithm we proposed is highly applicable to artificial broadleaf species recognizing.