Distinction and estimation of the area with conservation agriculture using SPOT 4 images

Abstract

The state of Guanajuato presents serious problems regarding the erosion of farming soils as a result of traditional preparation practices, and therefore, in the state of Guanajuato there are currently institutions that promote the adoption of conservation agriculture as an alternative to revert severe erosion. However, current statistics on the adoption of conservation agriculture in Mexico are not reliable, and when generating this information, the institutions that promote this practice could monitor and plan their activities. This investigation presents the results of a supervised classification of SPOT 4 satellite images, to distinguish and estimate the surface under conservation agriculture in an agricultural area of the state of Guanajuato. The methodology is based on a supervised classification analysis using the method of Mahalanobis and parallelepipeds. The analysis required the use of XS- SPOT4 and HRG2 SPOT5 images. A stratified-systematic field sampling was performed. The classification was evaluated using the matrix of confusion and the Kappa index. Of the plots verified on the field under conservation agriculture, 92.7% were identified correctly, along with 93.7% for conventional agriculture. The Kappa index calculated for Conservation Agriculture was 0.9208, and 0.9396 for conventional agriculture. The distinction and estimation of conservation agriculture were carried out for the spring-summer cycle 2010, and 16.1% was estimated for conservation agriculture, out of 262 104 ha that were classified. TheKappa index and the matrix of confusion showed that the classification supervised allowed a good separability between conservation and conventional practices.