Dear Dipankar Lahkar, usually the size and shape of the transect depends on the questions you want resposnder in the area. Ideally the transect should cover all environmental heterogeneity of the place. I believe the articles below can be easier to decide on the size and format. I hope that's help.

1) Gentry, A. H. 1982. Patterns of Neotropical plant species

diversity. Evol. Biol. 15:1-84.

2) Magnusson, W. E., A. P. Lima, R. Luizão, F. Luizão, F. R. C. Costa,

C. V. Castilho, AND V. F. Kinupp. 2005. RAPELD: A modification

of the Gentry method for biodiversity surveys in long-term

ecological research sites. Biota Neotropropica Jul/Dez 5. http://www.

biotaneotropica.org.br/v5n2/pt/fullpaper?bn01005022005±en.

3) Magnusson et al 2008. A Program for Monitoring Biological Diversity in the Amazon: An Alternative Perspective to Threat-based Monitoring. BIOTROPICA 40(4): 409–411. doi: 10.1111/j.1744-7429.2008.00427.x

Maybe you could find some problems with the distance sampling assumptions. You could use a bayesian approach to make the analysis to relax those assumptions.

You can read this paper:

http://izt.ciens.ucv.ve/ecologia/Archivos/ECO_POB%202009/ECOPO2_2009/Eguchi%20y%20Gerrodette%202009.pdf

Hope this helps

For concealing habitats I would use pellet group counts, applying Distance Sampling software. The size of the transect depends on your questions, type of habitat, resources available, etc. But I’ve been using 1-km line transects, distributed all over my study area, covering all habitat types and I found the results to be quite robust. Just be sure that all habitat types are covered by your transects.

A suggestion:

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0088459#pone-0088459-g002