In this paper, a fast and effective 3D reconstruction method for the growth of greenhouse tomato plant is proposed by
using real organ samples and a parametric L-system. By analyzing the stereo structure of tomato plant, we extracts rules
and parameters to assemble an L-system that is able to simulate the plant growth, and then the components of the L-system
are translated into plant organ entities via image processing and computer graphics techniques. This method can
efficiently and faithfully simulate the growing process of the greenhouse tomato plant.
Proc. SPIE. 9159, Sixth International Conference on Digital Image Processing (ICDIP 2014)
KEYWORDS: Image processing algorithms and systems, Agriculture, Digital image processing, Detection and tracking algorithms, Visualization, Databases, Image segmentation, Digital filtering, Reconstruction algorithms, Binary data
The automatic segmentation and recognition of greenhouse crop is an important aspect in digitized facility
agriculture. Crop stems are closely related with the growth of the crop. Meanwhile, they are also an important
physiological trait to identify the species of plants. For these reasons, this paper focuses on the digitization process to
collect and analysis stems of greenhouse plants (tomatoes). An algorithm for automatic stem detection and extraction is
proposed, based on a cheap and effective stereo vision system—Kinect. In order to demonstrate the usefulness and the
potential applicability of our algorithm, a virtual tomato plant, whose stems are rendered by segmented stem texture
samples, is reconstructed on OpenGL graphic platform.