The analysis of wheat appearance quality based on digital image processing

The appearance of wheat such as its shape and color is expected to be of important reference value in agricultural breeding and quality testing. Aiming at an objective and accurate analysis of wheat appearance quality, this article adopted median filter and iterative threshold segmentation to transform the original wheat images into binary images where single wheat seed is labeled subsequently to extract 7 morphological parameters as well as 6 color parameters. The experiment result shows that color parameters based on nonlinear conversion algorithms, in contrast with morphological parameters, are more suitable to identify germinated wheat and musty wheat which are common scenarios in wheat harvest and storage.