Automated Technique to Determine Key Growing Stages of Wheat for High-Throughput Field Phenotyping Using Computer Vision

Determine key growing stages information of wheat canopies is one of the most important phenotypic characteristics of wheats growth and development. Since monitoring growing stages is still visually performed by human, it is laborious task, very time-consuming and subjective. Thus, a requirement of non-invasive method which is capable of observing growth stage information automatically and continuously is essential. In this paper, we present an automated computer vision technique that can identify the critical growing stages of wheat such as wheatear emergence and flowering time based on images visual content taken from the Field Scanalyser at Rothamsted Research. The introduced approach is robust to complex environmental changes and outdoor invariants. The high accuracy of the method allows agronomist to identify the critical growth stages of wheat automatically. Furthermore, although the proposed method is applied on wheat, it can be used on other plants such as rice, maize, etc.