Image Target Extraction Method Based on ROI

A target extraction method based on region of interest is presented to solve the difficult problem of infrared image target identification technology.Firstly,the regions of interest are detected by using two-dimensional maximum entropy method,on the basis of this,the local classification method is applied to the local neighborhood of each pixel in the region of interest,in order to analyse the pixel gray characteristics;Secondly,the spatial position characteristics of every pixel point is builded by calculation of chamfer distance,and is combined with the gray characteristic;Finally,the new determination method of threshold value is defined to fetermine the ownership of every pixel point,to achieve the separation of background and extraction of target.The results of experiments indicate that the proposed method has a higher accuracy and stability comparing with the classic otsu method and maximum entropy method in infrared target extraction,it has a certain value for problems of low contrast and difficult recognition of infrared image.