Estimation of Branch Angle from 3D Point Cloud of Plants

Abstract

Measuring geometric features in plant specimens either quantitatively or qualitatively, is crucial for plant phenotyping. However, traditional measurement methods tend to be manual and can be tedious, or employ coarse 2D imaging techniques. Emerging 3D imaging technologies show much promise in capturing architectural complexity. However, automated 3D acquisition and accurate estimation of plant morphology for the construction of quantitative plant models remain largely aspiration. In this paper, we propose an approach for segmentation and angle estimation directly from dense 3D plant point clouds. Experimental results show that the approach is efficient and reliable, and appears to be a promising 3D acquisition and measurement solution to plant phenotyping for structural analysis and for building Functional-Structural Plant Models (FSPM)

Segmentation of tree seedling point clouds into elementary units

“…An approach analogous to that used here has recently been published by Lou et al (2015), although they work on point cloud data generated using a multi-view stereo imaging system. The methodology is similar: a graph is first constructed to build neighbourhood relationships between the plant's points, and next a spectral clustering approach is performed on this graph.…”

“…More general approaches such as expectation maximization could be used, but as k-means clustering they do not naturally benefit from the neighbourhood information (graph edges). Note that Lou et al (2015) merge neighbouring clusters with similar normals, but this may lead to under-segmentation since different elementary units (e.g. two leaves) may have similar normals.…”

“…ε and k are user-chosen parameters. ε-Neighbourhood is, for instance, used by Belton, Moncrieff, and Chapman (2013), whereas the k-Nearest- Neighbours method is used by Côté et al (2009) and Lou et al (2015). These methods are convenient provided the density of the point cloud is uniform, which is not the case for our TLS data.…”

“…It is used for example by Lou et al (2015). k-means clustering randomly selects k initial 'means' among the points, with k being a user-defined parameter.…”

“…For example, the use of different imaging techniques has been proposed for plant phenotyping; see Li, Zhang, and Huang (2014) for a review. The most popular imaging techniques are based on single-lens cameras (Quan et al 2006;Paproki et al 2012), time-of-flight cameras (Chéné et al 2012;Alenyà et al 2013;Chaivivatrakul et al 2014;Xia et al 2015), or multi-view stereo imaging systems (Golbach et al 2015;Lou et al 2015;Rose, Paulus, and Kuhlmann 2015). All these methods allow the reconstruction and measurement of single leaves, although some of them require manual interaction (Quan et al 2006;Golbach et al 2015;Rose, Paulus, and Kuhlmann 2015) or prior knowledge of the plant (Chéné et al 2012;Chaivivatrakul et al 2014).…”