Since the details in quasi-homogeneous region will be degraded from a conventional global image enhancement methods such as histogram equalization. This degradation causes from the saturation of the gray levels in equalization process. Then the local histogram equalization for each quasi-homogeneous region will be used in order to improve the details in the regions. To obtain the quasi-homogeneous regions the image can be segmented by using a watershed transform. The watershed transform is based on mathematical morphology; therefore the segmented regions can be effectively separated meaningful. Two adjacent regions in images which almost similar in gray levels can be separted by the watershed transform. In the color image enhancement three spectral bands are enhanced independently. Then all the three bands are assigned to each processed image in order to produce a color composite image. By the proposed algorithm the result images show the better perception on image details. Therefore the high efficiency of image classification can be obtained by using this color image.

This report aims to present the random noise elimination of image based on edge preservation. This solution is a pixel processing using 29 templates. Each template consists of the considered pixel and its neighbor pixels. To eliminate noises the smoothing process replaces the new pixel intensity into the same considered processing pixel by the mean intensity of neighbor pixels. The condition of the template selection is based on the minimum variance of templates. As the experimental results this solution brings the peak signal-to-noise ratio (PSNR) higher than Gaussian filter Tomita’s marks Fukuda’s marks and Nagao’s marks and also gains contrast difference.

The objective of this research is to proposed a semi-automatic method for mapping electric poles using the structure from motion technique or SFM. First an epipolar geometry for each image pair is estimated. Then the 3D structures are obtained using incremental SFM technique which will give us both 3D point cloud of the study area and camera parameters. Also vertical 2D lines are extracted from the input images. These 2D lines and the 3D structures are then integreated for estimating the positions of electric pole in 3D space. Finally these 3D electric poles are then geo-registered by estimating 3D rigid transformation from embeded coordinates in the input images.

Image enhancement is one of using in various digital signal processing areas. Advances in microcontrollers DSP boards and computers have been developed traditional algorithms to improve the quality of the resulting image. This paper aims to present the contrast enhancement using weighted bi-histogram equalisation. In additional this technique must use two weighted factors which is calculated by the histogram distribution ratio. Likewise an original image will be equalised by the modification of the probability density function of the gray levels. As the experimental results the contrasts of resulting images are improved for robot vision implement and human perception including absolute mean brightness error (AMBE) limitation.