In order to improve the situation of laser ranging signal extraction problem and degree of automation, the feature of laser ranging data are listed, considering these feature and combination of binary image analysis a new method extracting the signal from the laser ranging data that contain a lot of noise is proposed. Before removing the noise the original data should be mapped into a binary image matrix in the method. According to the density of signal is higher than noise, the first step is done to remove the noise in the image matrix. The second step is done to remove the noise based on the feature that the signal is similar to a line in the image. The result of the image which has been processed should be inversely mapped back. The experimental results show that it is a potential method which can extract signal automatically and replace artificial methods.