Automatic acquisition characteristic parameters of wheat ear based on machine vision

wheat ear characteristic parameters are important parameters for breeding and investigation of new wheat variety and could be used to estimate yield. To realize non-contact and accurate measurements of characteristic parameters of wheat ear, a computer vision measurement method based on mathematical morphology was proposed. The 5 ear traits, namely ear length spike shape, kernel top, spikelet number awn length and number were measured by image processing from 30 ears of six cultivars. The main methods include image segmentation algorithm, principal component analysis, template matching algorithm etc. Relative measurement errors were 4%, 2% and 6.2% respectively for ear length, awn length and number by image processing. The repeatability accuracy of spikelet number achieves ± 1. Spike shape keeps consistent with the results of visual observation. Image processing is a useful tool for extracting characteristic parameters from wheat ear, and will become more and more important in yield estimation, the new wheat variety DUS testing and breeding in the whole country.