Depth imaging-based detection of muskmelon plant for phenotyping in the greenhouse

As the rapid development of plant breeding and genomics, more effective and reliable phenotyping data has become the bottleneck for modern genetic crop improvement. In this article, we assess the potential of depth imaging system for phenotyping muskmelon. An original algorithm to detect muskmelon based on fusion images of RGB image and Depth image was proposed from complex backgrounds of greenhouse. Two types of features are extracted: 1) features from depth information; and 2) statistical features derived directly from RGB images. From muskmelon plant detected, various measurements of muskmelon phenotype involving for number of leaves, size of leave, size of fruits will be presented to demonstrate the practical interest of such imaging systems.