Development of an algorithm for online measurement of percent residue cover

To develop the prototype of an online-capable camera sensor for measuring percent residue cover, appropriate image acquisition equipment and exposure conditions were investigated and different image processing algorithms for segmenting images into residue and soil were written with the help of commercial software. A large number of observations can easily be obtained using an online sensor, and this makes evaluation of its correct functioning by means of visual standard methods for estimating residue cover questionable. Hence an independent automatable approach was developed based on the analysis of features extracted from grayscale histograms of the residue images, which is even faster than the classical segmentation approach. The sensor was mounted onto an all-terrain vehicle (ATV) and the residue cover of three test fields was analysed. The Pearson correlation between the two measurement approaches was 0.967 for the cover rate observations taken on the three fields.