Understanding the physical and especially mechanical properties of forest soils is very important in forest engineering operations including road construction and exploitation. Unconfined compressive strength (UCS) is one of the most important mechanical properties for evaluating the strength of adhesive soils in many engineering projects including forest roads. Measuring the USC of soil in the laboratory and in situ is difficult, time-consuming, and costly, so it is essential to explore a rapid, low-cost, and non-destructive method. The aim of this study was to estimate the UCS of forest roads using spectroscopy. The applied support vector machine (SVM) method determined the soil moisture classes (i.e., 14%, 25%, 31%, and 36%) with R² of 0.98. Two methods of partial least squares regression (PLSR) and the normal difference index (NDI) were used to estimate the UCS at different moisture classes. The results showed that the NDI performed better than the PLSR to estimate the USC at four different moisture classes of 14% (R² = 0.8, RMSE = 19.36, RPD = 2.61), 25% (R² = 0.78, RMSE = 16.06, RPD = 1.64), 31% (R² = 0.82, RMSE = 11.2, RPD = 2.28), and 36% (R² = 0.83, RMSE = 7.49, RPD = 2.41). Also, a sampling interval method was applied to reduce spectral dimensionality and processing time. The sampling interval of 16 nm was selected using a genetic algorithm. Finally, the UCS was estimated at maximum soil moisture using the simple regression model (SRM) based on the results of NDI. Based on this current study, it has been found that using spectroscopy to estimate the UCS of soil could be considered an alternative method for developmental operations in identifying the UCS of soil due to the costly and time-consuming conventional laboratory methods.

在森林工程操作（包括道路建设和开发）中，了解森林土壤的物理特性，尤其是机械特性非常重要。在包括森林公路在内的许多工程项目中，无侧限抗压强度（UCS）是评估粘合土壤强度的最重要的机械性能之一。在实验室和原位测量土壤的USC困难，费时且昂贵，因此探索一种快速，低成本且无损的方法至关重要。这项研究的目的是使用光谱法估算林道的UCS。应用支持向量机（SVM）方法使用R ²确定土壤湿度等级（即14％，25％，31％和36％）为0.98。采用两种方法的偏最小二乘回归（PLSR）和正态差异指数（NDI）来估算不同水分含量下的UCS。结果表明，在四种不同的水分含量分别为14％（R ² = 0.8，RMSE = 19.36，RPD = 2.61），25％（R ² = 0.78，RMSE = 16.06， RPD = 1.64），31％（R ² = 0.82，RMSE = 11.2，RPD = 2.28）和36％（R ²= 0.83，RMSE = 7.49，RPD = 2.41）。另外，采用了采样间隔方法来减少光谱维数和处理时间。使用遗传算法选择16 nm的采样间隔。最后，基于NDI的结果，使用简单回归模型（SRM）对UCS进行了土壤最大水分估算。基于当前的研究，已经发现，由于昂贵且耗时的常规实验室方法，使用光谱法估算土壤的UCS可以被认为是用于识别土壤UCS的开发操作的替代方法。