Quantitative detection of soybean rust using image processing techniques

Rust caused by Phakopsora pachyrhizi Syd. is a major constraint to soybean product in Asia. Early detection and possibilities of controlling plant diseases by the integration of several image processing methods has been the subject of extensive research. The main contribution of this paper is to present different methodologies for quantitatively detecting soybean rust at each stage of disease development, identify disease even before specific symptoms become visible and grade based on percentage of disease severity. Severity of rust infection levels at each stage of disease development was observed for 25 days on soybean leaf. Then color distribution and pixel relationship in rust infected leaf image was calculated based on global and local features for quantifying rust severity. Further, rust disease was categorized into grades based on infection severity levels and percentage disease index (PDI) was calculated. The maximum PDI of 95.5 was observed at 25th day and minimum PDI of 0.2 was observed at 6th day.