Object Oriented Multispectral Images Segmentation and Classification
Paper ID : 1407-IST
1Fardin Mirzapour *, 2Hassan Ghassemian
1Faculty of Electrical & Computer Engineering, Tarbiat Modares University
2Faculty of Electrical & Computer Engineering
Tarbiat Modares University
In this paper an object oriented method for on-line scene segmentation is developed. Normally, in remote sensing a scene is represented by pixel-oriented features. It is possible to reduce data redundancy by a segment-feature extraction process, where the segment-features, rather than the pixel-features, are used for multispectral scene representation. The algorithm partitions the observation space into a set of disjoint segments (called objects). Then, pixels belonging to each segment are represented by segment features. In this paper, an unsupervised seg-mentation algorithm based on statistical region merging (SRM) framework is pre-sented. Also a partial differential equations (PDE) algorithm is suggested as a pre-processing phase to improve the segmentation results. Illustrative examples are presented, and the performance of the features is investigated. Results show aver-age compression ratio about 31, the classification performance improvement for all classes, and less computational time consumption for classification.