METHOD FOR 3D SOIL HORIZONATION USING DIGITAL IMAGES

Abstract

We created a three-dimensional (3D) model of the spatial arrangement of soil horizons with broken boundaries using digital images. T is technique was tested on a Retisol - an Alf sol with a glossic horizon. Photographs were taken for 11 vertical sections (2.5 cm distance between sections) of the soil prof le for an area of 30 × 45 cm. Colorimetric accuracy of the images was tested against measurements of moist soil samples made with a portable spectrophotometer. T e selected best images were calibrated using an internal color calibration method for color correction. T e images were combined into a common array to build a 3D optical soil horizon map using the CIELAB color space. A protocol for processing the 3D soil images was created that showed 3D soil structure. It was found that the CIELAB color coordinates can be used to distinguish and delineate AE, E, and EB horizons. We then tested the method to assess soil carbon stocks and found that the stocks using the 3D model were 28% higher than when calculated using the 2D model. We conclude that the optical 3D mapping method can accurately represent the 3D structure and can be used to quantify soil horizon variations.