Plant location by computer vision used ratios of red, green and blue channels from colour images of the crop scene in order to reduce the effects of shadows following the method of Marchant and Onyango (2000). In the resulting mono-chrome images plant material appears bright against a dark soil background. An algorithm based on two-dimensional wavelets has been developed to locate individual crop plants. These wavelets defined by Eq. (1) provided a spatially localised means of extracting a periodic planting pattern based on individual plants and their near neighbours (Fig. 7).