Image analysis system to determine crop row and plant positions for an intra-row weeding machine

This paper describes computer vision methods to trace crop row positions and to locate the single crop plant positions in the rows. For the determination of the crop row a template fitting algorithm was developed. Detection of individual crop plant positions was based on a Fast Fourier Transformation (FFT). The results show that it is possible to detect the crop row and crop plant positions in a robust manner for different crops and different growth stages. On average less than 1% of the crop plants were not detected and less than 1% of the weed plants were classified as crop.