土壤结构的数字图像分析方法与指标

A digital image processing tool and indices for soil structure

自然地理过程的水、热应力及人为过程的机械应力可以促成土壤结构体的形成,然而土壤结构的分析方法与描述指标仍然较为局限,筛分法仅能提供土壤结构的尺度分布信息,感官评价法受主观影响也较大。本文探讨数字图像处理技术定量土壤结构并提供3个外观几何指标,分别从土壤结构尺度分布、矩形度、棱角性、形状指数4个方面评价黄棕壤块状结构体。结果表明自然地理过程作用下黄棕壤呈块状分布且随结构体尺度的增大,结构体的形状指数及棱角性呈增大的趋势,而矩形度则不断降低。所提供的数字图像分析方法快捷方便,较之筛分法能够提供更多的土壤结构信息,所得土壤结构尺度分布模型更加准确,本方法也可作为实时实地分析耕作过程土壤结构的变化。

Soil structure is a result from either the hydro-or thermo-stresses occurred in natural geological process, or mechanical stresses induced from anthropic activities. However, analytic tools and indices for soil structure are largely restricted. Traditional sieving analysis provides only size distribution of soil structures, while the visual assessment is prone to subjective errors. We investigated an image processing tool and proposed 3 indices for soil structure analysis.Structure of Hapludalf was assessed to provide its aggregate size distribution, rectangularity, angularity and form index.Results showed that under natural geological processes Hapludalf was appeared as blocky, in which form index and angularity increased with enlarged scale, while the rectangularity decreased. The proposed image processing tool was quick and efficient, providing more information than the traditional sieving method. Accuracy of the acquired soil structure size distribution was improved. The proposed method can be used for on-line and in-situ soil structure analysis.