Morphological and structural data acquisition for above-ground part of grapevine

Abstract: Three-dimensional (3D) reconstruction of aerial plants using in-site measurement morphological data, which is effective and highly realistic, is one of the main approaches to show the morphological diversity of diverse cultivars and intra specific differences caused by cultivation and environment factors. In order to satisfy the demand of digital plant modeling using in-site measurement morphological data combined with the advanced instruments and approaches of current plant morphological data acquisition, in this study, we made a systematic exposition of the 3D morphological data acquisition method of aerial grapevine. Firstly, a target plant of grapevine was selected according to the demand of modeling cultivar, tree structure and cultivation factors. Secondly, to simplify the complex structure of grapevine, the topological structure of target grapevine was subdivided into structure units which were identified by a naming strategy to distinguish organ type, growth order and inter-connected relationships. This gave a digital representation of the real target grapevine and facilitates subsequent digital operations. On the basis of digital representation, morphological data acquisition was conducted including four parts:1) The topological structure of the whole grapevine plant determined the reconstruction accuracy and sense of reality, hence to describe the 3D topological distribution of plant skeleton, 3D digital data in point skeleton form of the target plant was captured using 3D digitizer. This process had many difficulties actually, such as wide range, complex field environment, electromagnetic interference, etc. We used the FastScan digitizer and its probe combined with Polhemus Long Ranger calibration system to overcome these difficulties. 2) The DUS (Distinctness, Uniformity and Stability) information, which was used to describe the cultivar feature of plants, was captured randomly on target cultivar grapevines. DUS information reduced the data acquisition work on the basis that the measured morphological data could reflect cultivar characteristics statistically, especially providing guidance for choosing representative grapevine organs when acquiring 3D scanning and texture image data. In addition, DUS information provided a statistical description of organ parameters which were hardly to measure due to the huge amount of organs. 3) Organ morphological data was obtained, including 3D scanning and 2D texture image. The 3D point cloud obtained by Artec EVA and Spider 3D scanners was used to reconstruct geometric models of grapevine organs, where the organs were chosen according to the DUS information to be more representative. Geometric models of organs had a unique name consistent with the naming strategy as a structure unit. 4) Supplementary information such as current growth period, cultivation density, special treatment of target grapevine, consulted from the grapevine growers and experts, were recorded to describe the growth environment. To demonstrate the data acquisition approach, we acquired the morphological data of five different kinds of cultivar grapevines during the florescence to the first fruit-swell stage as the target plant, including a structure diagram of the grapevine; a morphological and texture DUS information; 3D point clouds with texture of the grapevine truck; 3D topological representation of the grapevine obtained by 3D digitizer; point clouds of leaves with texture; and point clouds of new shoot internode with texture. Combined with parameterized plant modeling technique and the measured morphological data, the geometric models of grapevine plant and organs were reconstructed with highly realistic. The models were used for 3D exhibition of the 11th international conference on grapevine breeding and genetics. Finally, a procedure of plant morphological data acquisition was extracted and discussed for other plants, mainly including three parts: target plant selection, digital representation of plant structure and data acquisition. The procedure could constraint the morphological data acquisition of plants more systematic and standardized, which has significance for further study and provides efficient and accurate data to support the research of digital plant.