Published Paper Details:

In obvious, the objective is to improve the performance of visual saliency aided optimal region growing for road region extraction from high spatial resolution multispectral satellite images. Satellite and aerial images are the most important available data sources for map generation and updating of available maps and they have highly improved in terms of spatial, spectral and temporal resolutions. In the proposed model satellite images is extracted using feature extraction such as color, intensity and orientation. Image transformation through Lifting Wavelet Transformation (LWT) is make use of three feature extractions then the image interpolation using bi cubic method is used for resize the extracted images and after applying this bi-cubic method the saliency map is located. The visual saliency Facilitated aided optimal region growing (VSFORG) is used for road region segmentation. In this segmentation different optimization techniques is applied for threshold optimization. In the result various evaluation metrics is analyzed mainly the accuracy for GWO is 82%, for ABC is 74% and for PSO is 64%. The proposed model is performed high when compared with other algorithms in the threshold optimization process.