Abstract: With the advances in automated production,manual operation decreasing fast in the production line,and machine vision come into being.It has high speed,precision,long running time and other characteristics.For the detection of the wounds of the bottom of the battery,we are currently using manual inspection,which has low efficiency and precision.This paper presents a method to detect the wound of broken skin of the battery bottom.We binarize the picture of the battery,and then to use morphological process.Finally we draw conclusion by counting the number of connected regions.This method does not use complex spectral analysis,and use morphological methods to eliminate noise,to achieve the purpose of discriminating wounds.This method greatly reduces the amount of computation to improve determination efficiency.