Vegetation indices derived from high-resolution airborne videography for precision crop management

This paper examines the potential of airborne videography as a remote sensing tool in a project aimed to develop a rapid and cost-effective system for assessing and monitoring the conditions of crops, pastures and native forest areas in the agricultural region of Western Australia. This implies a close examination of current techniques for low cost/high resolution data capture, and the requirements for image-based precision farming. The research approach comprises reviewing stress indicators that can be used as early signs of changes in the conditions of crops and pasture, and the identification of value-added products, such as vegetation indices, that can be used by farmers and land planners interested to know where field variations occur. Four vegetation indices, Namely Plant Pigment Ratio (PPR), Photosynthetic Vigour Ratio (PVR), the Normalized Difference Vegetation Index (NDVI) and the green NDVI (NDVIg) are tested. Experimental results related to the identification and mapping of crop density variations, crop types, and variations in crop conditions due to the presence of weeds and dead standing vegetation are discussed. Statistical analysis ( T -test and multiple comparisons procedure using the Fisher LSD test, at @ =0.05) suggest the PVR as the best index to detect relative variations (e.g. high, medium, low) in crop density, followed by the NDVI. Results also show the indices ability to separate canola from lupins and wheat. Additionally, the PPR appears to detect the presence of weeds ( Arctotheca calendula ) in pasture and cereal crops, while NDVIg and NDVI can only identify weeds in pasture. The NDVI and NDVIg appear as the more sensitive indices to detect the presence of dead standing vegetation (stubble) in pasture. The status of high-resolution satellite imagery, as compared to airborne videography, for crop management applications, and main findings in regard to the potential of video-based vegetation indices to support precision farming, are presented as well.