Journal of Forestry Research

, Volume 22, Issue 1, pp 99–105 | Cite as

Classification of features selected through Optimum Index Factor (OIF) for improving classification accuracy

Original Paper

Abstract

The present investigation was performed to determine if the features selected through Optimum Index Factor (OIF) could provide improved classification accuracy of the various categories on the satellite images of the individual years as well as stacked images of two different years as compared to all the features considered together. Further, in order to determine if there occurs increase in the classification accuracy of the different categories with corresponding increase in the OIF values of the features extracted from both the individual years’ and stacked images, we performed linear regression between the producer’s accuracy (PA) of the various categories with the OIF values of the different combinations of the features. The investigations demonstrated that there occurs significant improvement in the PA of two impervious categories viz. moderate built-up and low density built-up determined from the classification of the bands and principal components associated with the highest OIF value as compared to all the bands and principal components for both the individual years’ and stacked images respectively. Regression analyses exhibited positive trends between the regression coefficients and OIF values for the various categories determined for the individual years’ and stacked images respectively signifying the prevalence of direct relationship between the increase in the information content with corresponding increase in the OIF values. The research proved that features extracted through OIF from both the individual years’ and stacked images are capable of providing significantly improved PA as compared to all the features pooled together.

Key words

OIF supervised classification principal components band combinations 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Chavez PS, Berlin GL, Sowers B. 1982. Statistical method for selecting Landsat MSS ratios. Journal of Applied Photographic Engineering, 8: 23–30.Google Scholar
  2. Chavez PS, Guptill SC, Bowell JA. 1984. Image processing techniques for thematic mapper data. Proc. ASPRS-ACSM Tech. Paper, 2: 728–742.Google Scholar
  3. Jasinski MF. 1996. Estimation of sub pixel vegetation density of natural regions using satellite multispectral imagery. IEEE Trans. Geosci. Remote Sensing, 34: 804–813.CrossRefGoogle Scholar
  4. Palaniswami C, Upadhyay AK, Maheswarappa HP. 2006. Spectral mixture analysis for sub pixel classification of coconut. Current Science, 91: 12–25.Google Scholar
  5. Perumal K, Bhaskaran R. 2010. Supervised classification performance of multispectral images. Journal of Computing, 2(2): 124–129.Google Scholar
  6. Saha SK, Kudrat M. 1991. Selection of spectral band combination for land cover/land use classification using a brightness value overlapping’index (BVOI). Journal of the Indian Society of Remote Sensing, 19(3): 141–147.CrossRefGoogle Scholar
  7. Siljestrom PA, Moreno A, Vikgren K, Caceres LM. 2002. The application of selective principal components analysis (SPCA) to a thematic mapper (TM) image for the recognition of geomorphologic features configuration. International Journal of Remote Sensing, 18: 3843–3852.CrossRefGoogle Scholar
  8. Singh SK. 2006, Inside Jharkhand — 2nd edition. Ranchi: Crown Publications, India.Google Scholar

Copyright information

© Northeast Forestry University and Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  1. 1.Department of Remote SensingBirla Institute of Technology MesraRanchiIndia

Personalised recommendations