Fig. 1
From: PhenoTrack3D: an automatic high-throughput phenotyping pipeline to track maize organs over time
Overview of the 3D maize reconstruction and segmentation pipeline used prior tracking. A Maize plants grown in the PhenoArch high-throughput phenotyping platform. B Daily acquisition of 12 side-view RGB images. C Reconstruction of a 3D volume, using space-carving [16]. D 3D skeletonization of the reconstruction [16], and extension of the leaf tips. E Detection of the stem tip position (red box) using a deep-learning model. This position is used to segment the skeleton into stem (black), ligulated leaves (blue) and growing leaves (orange) organs. Ligulated leaves can be ordered topologically (numbers) by increasing insertion height, unlike growing leaves which all emerge from the same point