Data acquisition for 3D surface modelling of chilli plant by using close range photogrammetry for volume estimation

摘要

Besides reasonable price of software and instruments, the close-range photogrammetry (CRP) technique also considered as a simple and effective technique used to model almost any kind of objects in three-dimensional (3D) form. The model is constructed by matching a bunch of points marked in an image with its corresponding match in the other overlapping images. Once the task is finished, a 3D surface model is constructed by forming a triangulation network which developed by connecting the cloud points generated from a set of images of a particular object. In term of object that could be modelled, CRP techniques also applied in modelling not just non-living things but also living things which have matte surface and static position while image is captured. As the benefit of the CRP technique is fully noticed, this research paper is focused on how to acquire data for constructing 3D model of a living plant (i.e: chilli plant) so that it can be used to estimate the plant volume. Photo Modeler Scanner (PMS) software is used in this study to calibrate the camera used in image capturing, mark and match the points of object in the images and construct the 3D surface model of the plant. For the whole-plant volume estimation purpose, the raw 3D surface model constructed is processed first to remove noisy points available before the volume value is computed by the same software. The volume computed is stated as estimation only because it is difficult to build model that follows the complex design of the plant and apart from that, not every part of the plant can be captured in a fix distance. However, the volume value is good enough to be used in leaf area index computation which used currently by botanists and farmers in determine the right ratio of water, pesticides and fertilizers towards healthy and productive plants.