Segmentation processed images by radar SAR
Segmentation processed images by radar SAR
Hamid Alemi1
1) MSc computer software trends - Email:
IOI Code
:
XYCZ-AACHD
Publication :
International Conference on Engineering and Applied Science(iceasconf.com)
Abstract :
Three main specifications are involved in examination of image quality: resolution, contrast and image noise. Images taken by SAR radar have high noise levels. Such images are often difficult to classify. To separate different parts Pictures, Segmentation must be performed correctly and accurately. In this paper is proposed a method for optimal segmentation. To use deep belief networks is proposed combining genetic algorithm. The process will start by reading the radar images then deep belief network is used features the image elements and organized pixels in groups. Each of the group as a cluster center is used and the Center prepared for Genetic Algorithm Combined to select the optimal solution in image segmentation. This highly efficient combination of the results of this work indicate that improvement is very high. The proposed algorithm show Feasibility, robustness, accuracy and good efficiency for resolution of the images.
Keywords : image segmentation deep belief network noise hybrid genetic algorithm