Determination of Early Stage Corn Plant Height Using Stereo-vision

Periodic measurements of plant height could be used to calculate growth rate and have potential for characterizing spatial variation of plant response to inputs and field conditions. A corn plant height measurement system using stereo machine vision was developed. A test apparatus was developed, which consisted of an x-y translating sensor platform with a CCD camera and an imaging stage for camera calibration and plant image acquisition. The position of the camera was controlled precisely by a microcontroller and allowed images to be taken of the top view of the plant from a canonical stereo configuration. Corresponding plant edge points in two images were found using an area-matching algorithm after plant object segmentation. Plant height was estimated by triangulation using pixel disparity between two images. Average plant height was compared to manual measurement of maximum plant height. The stereo vision system estimated maximum plant height with a RMSE of 2.55 cm. Sources of error were investigated. These results demonstrate the potential of stereo vision for a vehicle-mounted crop height measurement system.