ABSTRACT
This chapter presents a comprehensive and critical survey of segmentation techniques developed for above-ground plant phenotyping applications. From a practitioner’s point of view, three main parameters, that play pivotal roles in the choice of segmentation algorithms, are: phenotyping platforms, imaging modalities, and phenotyping categories. The chapter focuses on segmentation algorithms using single images; segmentation based on multiple images. It presents the state-of-the-art traditional and learning-based segmentation techniques used for plant phenotyping. The chapter summarizes publicly available benchmark datasets, and discusses the potential open problems. It also summarizes the traditional and learning-based segmentation techniques used for holistic and component plant phenotyping analysis in both controlled environment and field-based phenotyping platforms. Color image segmentation is the process of partitioning an image into different regions, based on the color feature of the image pixels.
