This paper discusses the basic principles,operations and features of genetic algorithms,and takes the method of maximum classes square error as an example to describe the method and procedure about applying genetic algorithms in optimizing the selection of image threshold.

In the paper,firstly,using the classification theory in pattern recognition,approach of multithreshold stone image segmentation based on improved 2D class variance is presented.

A segmentation algorithm based on the technique of one-dimension maximal variance between-class and region growing method is proposed in this paper.

It was testified and compared with the Otsu method and the maximum entropy method.

Base on analyzing image segmentation methods now available,a method of flame image segmentation is put forward to solve gray-level histogram shape limit and unsteady segmentation effect of the Otsu method.

First the Otsu method was implemented to select a threshold in luminance field of the image automatically.

Application of the OTSU Method in Image Processing;

This paper proposes a rapid threshold method which combines the GCPSO with Otsu.

First, use OTSU method to make license plate image binarization, and then use genetic algorithm to map the entire license plate matching search feature.

At the same time, the method for selecting optimal threshold vector called extended class vector square error is given based on class square error (Os.