Image segmentation is crucial to object-oriented remote sensing imagery analysis. In this paper, a newly modified texture segmentation algorithm is proposed using spectral, shape and intensity features. This algorithm is a robust technique that can be applied directly to the color images. The image is pre-processed using Adaptive Switching Median Filter, which removes the impulse noises and keeping the fine details of the image intact in the most efficient manner. Also, the pre-processed image is smoothened using morphological operators, which reduces the false detection of abnormal cells. Then, the pre-processed image is transformed into HSV (Hue, Saturation and value) color space representation in order to analyze and establish a color contrast gradient. The multiscale morphological gradient in the intensity channel of the pre-processed image is obtained and multiplied with the color contrast gradient. The shape feature is extracted from the pre-processed image based on the descriptors such as compactness, convexness, rectangularity and eccentricity, moment invariants. Based on these spectral, shape and intensity features, markers are extracted for this image and given as input to the watershed algorithm which uses a Hill-Climbing approach to identify and label the neighborhood pixels. This algorithm may reduce the computational complexity by avoiding the process of computing lower-complete image.

1. Willhauck. G. 2000 classification techniques and standard image analysis for the use of change detection between SPOT multispect Int. Arch.Photogramm. Remote Sens., vol.XXXIII, pp.35 42, Supplement B3.
2. Comaniciu. D and Meer. P May 2002 Mean shift: A IEEE Trans. Pattern Anal. Mach. Intell., vol.24, no.5, pp.603 619.
3. Rother. C Kolmogorov. V and Blake. A, Aug. 2004 ACM Trans. Graph., vol.23, no.3, pp.309 314.
4. Shafarenko. L Petrou. M and Kittler. J, Nov. 1997 rshed segmentation of randomly IEEE Trans. Image Process., vol.6, no.11, pp.1530 1544.
5. Zhao. Y, Zhang. L, Li. P and B. Huang, May 2007, improved Gaussian Markov random-field-based texture IEEE Trans. Geosci. Remote Sens., vol.45, no.5, pp.1458 1468.
6. Hsin H.-C. , Jul. 2000 IEEE Trans. Image Process., vol.9, no.7, pp.1299 1302.
7. Randen. T and Husoy. J. H, Apr.1999 segmentation using filters with optimized energy IEEE Trans. Image Process., vol.8, no.4, pp.571 582.
8. Zhang. X Jiao.L, and Liu. F, Jul. 2008 clustering ensemble applied to SAR image IEEE Trans. Geosci. Remote Sens., vol.46, no.7, pp.2126 2136.
9. Chen. J, Pappas. T, Mojsilovic. A, and B. Rogowitz, Oct. 2005 texture image IEEE Trans. ImageProcess., vol.14, no.10, pp.1 13.
10. Deng. Y and Manjunath. B, Aug. 2001 ervised segmentation of spectral texture regions in images and IEEE Trans. Pattern Anal. Mach. Intell., vol.23, no.8, pp.800 810.
11. Akçay. H. G and Aksoy. S, Jul. 2008 detection of geospatial objects using multiple hierarchical segmentati IEEE Trans. Geosci. Remote Sens., vol.46, no.7, pp.2097 2111.
12. Nan Li, Hong Huo, and Tao Fang, Jul. 2010 Texture-Preceded Segmentation Algorithm for High- IEEE Trans. Geosci.. Remote Sens., vol.48, no.7, pp.2818 2828.
13. Changren Zhu, Hui Zhou, Runsheng Wang, and Jun Guo, Sep. 2010 Detection from Spaceborne Optical Image Based on IEEE Trans. Geosci. Remote Sens., vol.48, no.9, pp.3446-3456.
14. Moga, A.; Cramariuc, B.; Gabbouj, M., Apr. 1995 efficient watershed segmentation algorithm suitable for IEEE Trans. Image Process., vol. 2, no.8, pp.101-104.

Color contrast gradient multi-scale morphological gradient shape feature marker texture segmentation