Machine Vision System for 3D Plant Phenotyping

Abstract

Machine vision for plant phenotyping is an emerging research area for producing high throughput in agriculture and crop science applications. Since 2D based approaches have their inherent limitations, 3D plant analysis is becoming state of the art for current phenotyping technologies. We present an automated system for analyzing plant growth in indoor conditions. A gantry robot system is used to perform scanning tasks in an automated manner throughout the lifetime of the plant. A 3D laser scanner mounted as the robot's payload captures the surface point cloud data of the plant from multiple views. The plant is monitored from the vegetative to reproductive stages in light/dark cycles inside a controllable growth chamber. An efficient 3D reconstruction algorithm is used, by which multiple scans are aligned together to obtain a 3D mesh of the plant, followed by surface area and volume computations. The whole system, including the programmable growth chamber, robot, scanner, data transfer and analysis is fully automated in such a way that a naive user can, in theory, start the system with a mouse click and get back the growth analysis results at the end of the lifetime of the plant with no intermediate intervention. As evidence of its functionality, we show and analyze quantitative results of the rhythmic growth patterns of the dicot Arabidopsis thaliana(L.), and the monocot barley (Hordeum vulgare L.) plants under their diurnal light/dark cycles.

Sensing Technologies for Precision Phenotyping in Vegetable Crops: Current Status and Future Challenges

“…Imaging techniques indeed represent, among others, a large portion of PRS-based measurements in phenotyping, and their applications cover a wide range of different types of observations for the estimation of many biometric and physiological parameters [14,15,23]. Imaging phenotyping can be accomplished through the simple acquisition of 2D images or by 3D models generated by multi-perspective acquisitions [24] and 3D images created by laser-scanner imagers [25,26]. With respect to the electromagnetic spectrum, measurements can be done in the range of ultraviolet (UV), visible (VIS), near-infrared (NIR), and infrared (IR) radiation [11], each providing different information.…”

“…However, 2D images present many limitations, especially when used for plants that have a high degree of structure complexity, therefore, 3D images are preferred [26]. The use of stereo camera rigs and the analysis by computer programs of images taken by multiple angulations allow drawing sophisticated models for the reconstruction of plant structures in 3D [24,34].…”

“…LiDAR technologies consist of an active laser sensor providing direct measurements of canopy architecture and organ distribution [25] for the estimation of plant volume, LAI, and biomass [25,35,37,38], thus allowing plant growth analyses from the vegetative to reproductive stages [26,39,40]. Other techniques include time-of-flight cameras and ultrasonic sensors, reviewed in [14].…”

“…The interaction between radiation and plant leaves-in terms of spectral reflectance, absorbance, and transmittance-can be defined as a passive process in which incoming electromagnetic radiation is affected by the relative quantity, but not in the wavelength form (as occurs, for example, in the fluorescence phenomenon). Reflectance, absorbance, and transmittance measurements can be related to plant structural and chemical characteristics [31,[41][42][43], thereby allowing the assessment of plant water and nutrient status [31,43,44] and photosynthetic activity, the detection of biotic and abiotic stress [25,27], and the evaluation of different plant physiological statuses [26,27,45]. In fact, reflectance, absorbance, and transmittance of electromagnetic radiation are influenced by plant tissue morphology and elemental and molecular composition [15].…”

PI-Plat: A high-resolution image-based 3D reconstruction method to estimate growth dynamics of rice inflorescence traits

“…100First, a rice panicle is often occluded by other plant components such as leaves and other panicles. 101Therefore, the existing solutions by moving cameras [42] are not entirely suitable to generate un-occluded 102 images for a panicle. Second, a panicle is non-rigid and typically is not located in the center of a plant, 103making it difficult to apply the existing solutions based on plant rotation [42].…”