Document Type : Original Article
Authors
1 Assistant Professor
2 Researcher of Agricultural Engineering Research Department, Kerman Agricultural and Resource Research and Education Center, AREEO, Kerman, Iran.
Abstract
Crop residue management is very important in farmlands. Conservatiion and proper management of residues, improves soil structure, retain moisture and reduces soil erosion. Whereas, crop residue burning converts organic material into ash, increases soil erosion and moisture loss from feild. The aim of present study was to research for accurate, fast and inexpensive methods for monitoring farms where crop residues are burned. keeping this in view, the potential of Landsat 8 sensor local data for monitoring residue burning was evaluated, using three classification methods including; supervised classification, unsupervised classification and detection of changes. Total number of 120 farms with 4 different surface coverage namely: plant residue, soil, green plant and residue ash were considered. Burned field’s location and their area were determined through satellite image with tree methods and their results were compared with field results. The results showed that due to successive changes in surface conditions of experimental farms, between two satellite imagery, such as tillage, seed planting and planted crop emergence, the satellite imagery could not be monitor the burned farms appropriately. Location and estimation of burned farm area by supervised classification was done with high accuracy. Overall classification accuracy of supervised classification method was 96.6, kappa coefficient was 0.93 and R2 was 0.92. Although by the unsupervised classification method some burned farms were separated, but overall classification accuracy and kappa coefficient of this method was low (71.6 and 0.61 respectively). Finally and based on the results it can be suggested that, supervised classification was chosen for farms remote sensing, where crop residues are burned.
Keywords
)