Precise Measurement of Residue Cover by Means of Image Processing Techniques

Abstract

Determination of crop residues on the ground is one of the basic requisites in many agricultural researches such as estimating the potential of
different conservation tillage tools for covering the crop residue. There are several indices for estimation the residue covers while most of them
are not only laborious and time consuming but also impressed by human errors. In this study a method has been developed which can be used to
measure the percentage of the residue covers based on image processing techniques. In as much as crop residues have various colours, three species
of crop residues were considered in this study including corn, wheat and canola residues. Colour spaces HSV, and RGB were used to investigate
the highest contrast between the residue and soil. The optimum combination of colour components in RGB colour space which can discriminate the
residues from the soil was also determined. Misclassification rate of the residues as the soil was measured less than 2.3%. Weight of the residues
per unit area was collated versus area residue covers gained from the image based technique. Results showed a good correlation between the corn
and canola residues determined from the images and their weights while for wheat it was depended on the density of the residues on the ground. It
was concluded that this method is practical for fast and convenient determination of residues on the ground.