An Image-Based Approach for Loamy Soil Dryness Classification Using SVM and Hybrid Features

Correctly identifying the moisture and dryness of loamy soil is essential to maintain agricultural productivity. Since loamy soil is favorable for crop root growth, its physical and moisture properties play an important role in determining irrigation management, crop planning, and soil development strategies. In conventional methods, soil moisture and dryness are usually determined by cutting or lifting samples from the soil and transporting them to the laboratory, which directly interferes with the natural structure and condition of the soil. Such methods are considered invasive, time-consuming, and relatively expensive. In order to overcome these limitations, in this study, we proposed a fast, non-invasive, and image-based method for predicting the dryness level of loamy soil, where hybrid image features and a support vector machine (SVM) classifier are used. In the proposed method, a comprehensive feature vector is formed by adding color features, Local Binary Pattern (LBP), Haralick texture features, and Histogram of Oriented Gradients (HOG) features extracts from soil images. Subsequently, different SVM models with linear, Radial Basis Function (RBF), and polynomial kernels were trained and evaluated using these feature vectors to classify the dryness of loamy soil into five categories: very dry, dry, moist, wet, and very wet. Experimental results indicate that the proposed model achieved high accuracy, precision, recall, and F1-score, which proves the effectiveness of the hybrid features and SVM kernels used, the stability of the model, and its applicability in real situations. Overall, the proposed image-based non-invasive method can be considered as a fast, cost-effective, and practical alternative for assessing the dryness of loamy soil by reducing the reliance on conventional laboratory-based techniques.