Remote Sensing Image Retrieval using Semantic Mining
Deepti Jhaman Punjabi1, Ajitkumar Khachane2, Ranjana Gite3
1Deepti Jhaman Punjabi, Student, Department of EXTC, Vidyalankar Institute of Technology, Wadala, Mumbai (Maharashtra), India.
2Ajitkumar Khachane, Associate Professor, Department of EXTC, Vidyalankar Institute of Technology, Wadala, Mumbai (Maharashtra), India.
3Ranjana Gite, Associate Professor, Department of EXTC, Vidyalankar Institute of Technology, Wadala, Mumbai (Maharashtra), India.
Manuscript received on 17 October 2014 | Revised Manuscript received on 24 October 2014 | Manuscript Published on 30 October 2014 | PP: 21-24 | Volume-4 Issue-5, October 2014 | Retrieval Number: E1814104514/14©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Understanding of images continues to be one of the most exciting and rapidly-growing research areas in various fields of technology. The recent advancements in hardware and telecommunication technologies like satellite communication in combination with the ongoing web proliferation have boosted growth of digital visual content on a large scale. However, this rate of growth has not been matched by the simultaneous improvement of technologies to support efficient image analysis and their retrieval. As a result, the overflow of available visual content resulted in large number of users facing hindrance in accessing information of the appropriate visual content. Moreover, with the immense number of diverse application areas that have emerged, which rely solely on image processing systems, has further revealed the tremendous potential for effective use of visual content through intelligent analysis. Better access to image databases, enhanced surveillance and authentication support systems, content filtering, adaptation and transcoding services, improved human and computer interaction, etc. are among the several application fields that can benefit from semantic image analysis or semantic mining. In this, images from desired database have been subjected to various steps involved in processing of images like pre-processing, segmentation, region level feature extraction and semantic mining. Satellite images are used to monitor the remotely sensed geographic area under consideration. Pre-processing involves steps where low level features are easily obtained using content based image retrieval scheme. Semantic mining technique is used to obtain other high level features for better image retrieval. Furthermore, region based segmentation allows systematic decoding of visual information and quantization based on different color intensities involved in the image. In this segmentation is performed based on the proposed JSEG (J Segmentation) algorithm. A probabilistic method will be used to mine the relationship among semantic features, regions, and images for region based feature extraction. Finally the Expectation Maximization method is used to analyze the relationship and extract the latent semantic concepts. This involves implementation of this approach on a dataset consisting of thousands of satellite images to obtain a high retrieval precision, thus solving our purpose.
Keywords: Segmentation, Image Retrieval, Object-Based Image Analysis, Remote Sensing (RS) Image.
Scope of the Article: Remote Sensing, GIS and GPS