Remote sensing recognition of paddy waterlogging using change vector analysis model

Abstract

It is significant to monitor crop waterlogging range timely and correctly for later field management, agricultural insurance and yield prediction. The change regularity of paddy injured growth was analyzed and the sensitive parameters of growth stress were screened. The model of identifying the paddy waterlogging range based on change vector analysis (CVA) was developed by using the HJ-1/2 CCD images around waterlogging. At last the accuracy of the developed model was evaluated by in-situ sample data. Results showed that the waterlogged paddy mainly scattered around the Huaihe River system. The spatial distribution pattern was in conformity with the occurrence tendency of paddy waterlogging provided by local agricultural department as a whole. By evaluating the accuracy of the model with in-situ samples, the overall accuracy of the model developed in the study reached 87.5%, while the Kappa coefficient reached 0.737. The change vector analysis model could identify the waterlogged paddy and normal paddy correctly and efficient in Huaihe River Basin. © 2013 IFIP International Federation for Information Processing.