Abstract:
The objectives of this study is to identify the idle land found in agricultural zones of Pan Thong district, Chonburi Province, Thailand. The proposed remote sensing data for identifying the Idle lands is the use of the NDVI Time-Series. The identification process is based on the users specific thresholds. It was found in the experiment that Maximum Likelihood Classification gained only 38% of the overall accuracy, On the other hand, The K-Means based method that was assigned the number of clustering classes to 20, 40, 60, 80 and 100 classes gained higher overall accuracies as the best classification accuracy was 70% (i.e., the results of the 40-class K-Means classification). The classification results confirmed that the proposed method can be used for identifying the idle lands of the study area. It is hope that the proposed methodology could be applied to other study areas of similar agricultural patterns. The future study will be the use of finer NDVI-Time Series, the used of other vegetation indices, and the addition of the ancillary data such as RADAR images.