Diagnosing method for plant growth using a 3D depth sensor
To simultaneously increase yield in horticulture, it is necessary to control the environment in the greenhouse while continuously monitoring plant growth. Light-intercepting characteristics evaluated from leaf area is used as a plant growth index. Therefore, we tested whether it is possible to measure leaf area nondestructively using a 3D depth sensor, Microsoft Kinect. The Kinect was developed to control video games through gestures and spoken commands. Information collected with a visible-light camera, an infrared-light camera for measuring distance, an infrared projector, and a microphone array is transferred to a Windows PC. We wrote a leaf-area calculation program in C# and built it in the Kinect for Microsoft Software Development Kit. Comparing the actual and estimated areas of plastic plates and using a linear approximation method gave an estimation accuracy of R2=0.99. We placed the Kinect above a tomato canopy in a greenhouse and measured the leaf area of two cultivars, 'Anoukou 9' and 'Managua'. We found that the spatial distribution of leaves was markedly different between cultivars, and that the Kinect calculated the sunlit leaf area. By comparing the leaf area and LAI obtained with the Kinect and the measurements of the vertical change in light intensity within the plant canopy, it was possible to describe the light-intercepting characteristics of each cultivar.
Umeda, H., Mochizuki, Y., Saito, T., Higashide, T. and Iwasaki, Y. (2018). Diagnosing method for plant growth using a 3D depth sensor. Acta Hortic. 1227, 631-636
plant growth, leaf area, 3D depth, image, non-destructive measurement, environment control