Abstract: A detection way for head and tail of Carya cathayensis based on the machine vision and image characteristics were proposed in order to adjust the best position for crumbling and improve the shell crumbling quality of Carya cathayensis during artificial processing. On the basis of mathematical morphology, the characteristics of Carya cathayensis, such as shape, texture, size, were established firstly, and a mathematics model was built according to the size and texture of Carya cathayensis; then the head and tail of Carya cathayensis were detected, and the suitable position for shell crumbling was determined to provide reference for the automatic shell remove of Carya cathayensis in future.