Development of a web‐based disease monitoring system for wheat crop
Loading...
Date
2016-03
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)
Abstract
Present study has been conducted in US Nagar district of Uttarakhand to monitor yellow rust disease in wheat using web based system. To monitor yellow rust disease in wheat, a web based system was developed using five step approach: a) discrimination of wheat and development of normal growth profile of the crop by employing remote sensing technique, b) identification of the incidence of yellow rust by constructing normal growth profile and identifying dips in the growth profile together with suitable weather conditions, c) computing inter annual deviation in mean NDVI from normal value to quantify the effect of yellow rust on growth of wheat d) developing a spectral/spectro-meteorological model for predicting and estimating wheat disease, and e) development of a web-based disease monitoring system. LULC map was generated using ENVI-4.8 image processing software for analyzing area under different entities with greater emphasis on crop area. This was achieved through the use of LANDSAT-TM/ETM+/OLI images of the years 2005-06 to 2014-15. Atmospheric correction of the images was carried out using Quick Atmospheric Correction (QuAC) technique in order to know the exact reflectivity of an entity in absence of atmosphere. Ground truth of wheat field for the month of February and March (2013-2015) was carried out for identification of the crops and wheat fields. QGIS software was used for digitization of district boundary of Udham Singh Nagar. Maximum likelihood classifier was used to generate LULC map and to discriminate wheat, sugarcane and mustard crop. SPSS software was used for the development of meteorological/spectro-meteorological model. The spectral reflectance decreases significantly with the increasing severity level in near-infrared and increases in blue and red bands. As a result the value of NDVI decreases with increasing severity of the disease. The normal growth profile of wheat crop was generated from 1st of November to 30th April by using mean NDVIs derived through SPOT data. The mean NDVI was calculated for each year and disease impact was analyzed using observed and predicted NDVImean. As observations on disease severity are not available, a disease severity index was developed for 2005-06 to 2014-15 by comparing the projected value of dip NDVI and the previous decadal NDVI. The disease severity index was used for the development of multiple regression meteorological models considering the weather parameters. Two meteorological models were developed and among them the performance of model 2 was found to be the best in wheat (R2=0.69). A spectro-meteorological model was also developed using remote sensing derived index of January (at 10 days interval) and the decadal meteorological parameter. The value of coefficient of determination for model 3 was 0.91, which suggest that remote sensing based model could prove to be an important tool foretelling yellow rust disease in wheat. Finally a web-based yellow rust disease monitoring system was developed using output of spectro-meteorological model 3 and converting them in vector format for all ten years. A simple one click on map displays a pop up to show disease severity level in per cent along with its management practices.
Description
Keywords
null