Optimization methods on dynamic monitoring of mineral reserves for open pit mine based on UAV oblique photogrammetry

The dynamic monitoring of mineral reserves for open pit mine by oblique photogrammetry is challenged by data-matching accuracy, illumination and terrain depression. In this paper, an object-level change detection method based on neighborhood windows is proposed to improve the accuracy of change detection in open pit mines. On this basis, a mining volume calculation method that considers mine type is established to solve the problem of lack of volume calculation caused by the depression of the mined terrain. A case study of eight open pit mines in the Shanxi Province indicated that the Precision (correctness), Recall (completeness) and F1-score for change detection were up to 92.38%, 87.22% and 89.73%, respectively. The relative errors in the calculated mined volume were between 0.38% and 0.77%. Our study provides the basis for creating an automated, highly robust workflows suited for monitoring mineral resource using 3D models acquired from UAV oblique photogrammetry.