The use of reflectance sensors to quantify crop development and weed competition

Abstract

This study was undertaken to examine a reflectance based system using modified WeedSeeker sensors as a means to non-destructively measure crop development. Four crops, wheat, canola, chickpea and fababean which have different canopy architecture were selected. The work was done in 2002 (a drought year) and 2003 (a good year) which provided contrasting seasonal conditions. Crop analysis using reflectance sensors accurately predicted biomass up to 1000 kg ha-¹ for all species. This was equivalent to a leaf are index of 1. Therefore the technique has the potential to rapidly and non-destructively determine early season crop development which can be used in a wide range of research and applied situations. Leaf area index increased with time and reached a peak between 110 and 120 days after solving. It then declined as the older leaves fell away. Reflectance increased with time and followed exactly the same pattern as leaf area index. Biomass increased over time and reached an asymptote after the peak in the leaf area index. Crop biomass was accurately predicted by reflectance measurements between the ranges of 0 to 1000 kg ha-¹. There was a greater similarity in the shape between the reflectance versus time curve and leaf area index versus time curve for canola, chickpea and fababean than there was for wheat. Leaf area index was found to be linearly related to reflectance for canola, chickpea, fababean and wheat. The leaf area index of the crops grown in 32 cm rows was generally higher than those grown in 64 cm rows. Row spacing needs to be added as a factor to the model to allow for the significant difference between the reflectance of narrow and wide rows. A different calibration equation would be required for each crop type and row spacing. The study also included examination of two applications: 1. Predicting weed competition. 2. The difference in the reflectance of chickpea varieties to weed competition.