Threshold infrared image segmentation based on improved genetic algorithm

As to the drawbacks of the Ostu method and Maximum Entropy Image Segmentation method, this paper proposes the threshold segmentation of infrared image based on an improved genetic algorithm. So it takes the Ostu of the image as the proper function, considering the pixel distribution between the aim and the background of the infrared image as well as the noise existing in the image. With the combination of this improved genetic algorithm method and maximum entropy image segmentation method, the optimal threshold value is obtained. The emulation experiment proves that this method is much more effective than the traditional genetic algorithm and the infrared image is better segmented in the end.