Automated Identification And Counting Of Pests In The Paddy Fields Using Image Analysis

A digital image analysis (DIA) algorithm based on fuzzy logic was developed using digital values of color, shapes and texture features to identify pests from images captured from a paddy field. Images were acquired under natural lighting using digital cameras and analog camera. Six pests species commonly found in the study site, Sawah Sempadan, Malaysia paddy field, namely rice leaffolder (Marasmia Patnalis), rice skipper (Pelopidas Mathias), rice leaf-butterfly (Melanitis Leda), Malayan black rice-bug (Scotinophara Coarctata), seedbugs (Pachybrachius Pallicorais) and rice butterfly (Abisara Saturate Kausambioides) were selected for this study. Image processing software, eCognition was used for discriminating analysis. Protocol that recorded all the procedure involved in the analysis process was built to automatically identify and count the total number of pests in the image. A satisfactory result of classification and counting of the pest was obtained from the image analysis. 100% accuracy was obtained for pest extraction and classification. There is 33.33% accuracy of the automation process in identification and counting of the pests obtained from the protocol built without the need of refinement. However an accuracy of 100% for automation process was obtained in identification and counting of the pests after the refinement of the protocol. The precision analysis system was capable of detecting pests from paddy plants images quickly.