Automatic identification of two growth stages for rapeseed plant: Three leaf and four leaf stage

Abstract

As an indispensable technology of agricultural development in the world, digital agriculture is the coalition of agriculture and modern information technology together with artificial intelligence technology. As one of the most important aspects of digital agriculture, crop growth monitoring requires for non-destructive real-time and accurate access to plant growth information in order to guide fine management of crop. Among them, monitoring of crop in different periods is an important component of plant growth monitoring. In this study, a computer pattern recognition method - ASM (Active shape model) is employed to automatically identify whether the rapeseed plant has reached three-leaf stage and four-leaf stage or not. Frist, the rapeseed plant blades are manually marked on the training samples. Then, all the blade models are aligned to obtain an averaged shape model, which is utilized as the geometric shape model. Finally, the averaged model is employed to search in the rapeseed plant image. Once the matched geometric model is found in the rapeseed plant image, the conclusion that the rapeseed plant has reached three-leaf stage and four-leaf stage will be drawn. Experiments are conducted on real images and the proposed method produce similar observation results comparing with the manual observation method. Therefore, the automated identification method can meet the demand for practical observation needed for agronomic modeling and in triggering action alerts to farmers.