基于UAV摄影测量技术的开采沉陷全盆地建模和求参

以内蒙古唐家会矿区为研究对象,获取该地区2020年8月与2021年3月无人机摄影影像数据,并制作生成DEM,将2期DEM数据相减获取该地区下沉盆地,用BP神经网络算法作去噪处理,并对比不同去噪方法的去噪效果;利用全盆地下沉数据,融合模拟退火算法(SA)与概率积分参数反演方法,求出该下沉盆地下沉系数与主要影响角正切;利用该参数模拟下沉盆地,计算出测量中误差为589 mm,占最大下沉值8.1%;最后对参数作抗差分析,在测量中误差占(1%~10%)最大下沉值时,求参结果可靠。结果表明:BP神经网络算法能够有效去除盆地内噪点,提高下沉盆地的精度,基于SA和矿区全盆地数据能够有效求取概率积分参数,弥补无人机精度不高带来的影响。

Whole basin modeling and parameter inversion of mining subsidence based on UAV photogrammetry technology

Taking Tangjiahui Mining Area in Inner Mongolia as the research object, the UAV photographic image data of August2020 and March 2021 in this area were obtained, and DEM was produced. The subsidence basin in this area was obtained by subtracting the DEM data, and the denoising effects of different denoising methods were compared with MATLAB software. Based on the subsidence data of the whole basin, the subsidence coefficient and the main influence tangent of the subsidence basin are obtained by using the probability integral parameter inversion with method of simulated annealing(SA). Using this parameter to simulate the subsidence basin, it is calculated that the measurement error is 589 mm, accounting for 8.1% of the maximum subsidence value. Finally, the robust analysis of the parameters is made, and when the error in the measurement accounts for(1%to 10%) the maximum subsidence value, the result of parameter calculation is reliable. The results show that the BP neural network algorithm can effectively remove the noise in the basin and improve the accuracy of the subsidence basin. Based on SA and the data of the whole basin in the mining area, the probability integral parameters can be obtained effectively, which can compensate for the influence of the low accuracy of UAV photogrammetry technology.