Abstract: BackgroundAccurate and rapid monitoring of vegetation cover in mountainous mining areas is essential for ecological research and soil and water conservation. Vegetation cover is a key indicator of ecosystem health and stability, especially in mining regions where human activities can significantly impact the environment. The Hutubi mining area in Xinjiang, China, has experienced extensive mining activities, leading to severe vegetation degradation and soil erosion. To address these challenges, it is crucial to develop an efficient and precise method for monitoring vegetation cover. This study aims to investigate the applicability and accuracy of various visible light vegetation indices derived from UAV imagery for estimating vegetation cover in this region.MethodsThis study integrates UAV-based visible light imagery and field survey data to evaluate the performance of 14 vegetation indices in estimating vegetation cover across different altitudinal gradients in the Hutubi mining area. Field surveys were conducted to collect ground truth data, which were used to validate the UAV-derived vegetation cover estimates. Grey relational analysis was employed to screen key variables, and the Ostu threshold method was used to calculate vegetation cover. Accuracy assessments were performed across various altitudinal gradients to evaluate the performance of different vegetation indices.ResultsThe results indicate that the applicability of vegetation indices varies with different levels of vegetation cover. The EXG index demonstrated superior accuracy in medium and high cover areas, with an accuracy (acc) exceeding 85% and a root mean square error (RMSE) of less than 0.5, significantly outperforming the other indices. In contrast, the identification of vegetation indices in medium cover areas revealed certain errors, necessitating further optimization through the integration of more topographical factors. The vegetation cover estimation method based on UAV visible light imagery offers the advantages of low cost, high efficiency, and strong spatial continuity. However, some misclassification and omission still occur in medium cover areas. The study also found that most indices achieved estimation accuracies above 70% in low and high cover areas, while the accuracy in medium cover areas was relatively lower.ConclusionsThis study provides reliable technical support for rapid vegetation monitoring and management in the Hutubi mining area. It offers important references for the application of UAV visible light imagery in vegetation monitoring in mountainous mining areas. The findings suggest that while UAV-based visible light imagery is effective for vegetation cover estimation, further improvements could be achieved by incorporating multispectral imagery or machine learning methods to enhance accuracy. This study addresses the critical need for accurate vegetation cover monitoring in mining regions and contributes to the broader goals of soil and water conservation and ecological restoration.