华中科技大学学报（自然科学版）2025，Vol.53Issue(3)：23-30,8.DOI:10.13245/j.hust.250235
基于无人机遥感平台的砾石自动筛分方法
Automatic gravel sieving method based on UAV remote sensing platform
摘要
Abstract
Aiming at the problem of measuring the particle size distribution of gravel beach particles,an automatic gravel sieving method(GAS)based on an unmanned aerial remote sensing platform was presented.First,a calibrated sphere of known size was suspended by a unmanned aerial vehicle(UAV)for remote image acquisition.Second,the maximum search range of gravel pixel patches was determined based on the convex features of the gravel image geometry.Third,the gravel pixel patches were searched and judged based on the characteristics of gravel in the color space.Fourth,the generation method of gravel images was determined based on the end method of pixel patch search.Fifth,the grain size was measured manually by applying ImageJ.The results show that the correlation coefficients between the results obtained by the GAS method and the manual measurements are 96.3%～99.6%,and the relative errors are 4.5%～29.7%.This result indicates that the GAS method has high accuracy and reliability in particle size measurement.In addition,the results were compared with the automatic image filtering method(AIF)and the BASEGRAIN method.The results show that the relative errors of the cumulative curves generated by the GAS method are 4.5%～29.7%.In comparison,the relative errors of the cumulative curves generated by the AIF and BASEGRAIN methods are 10.9%～84.6%and 23.1%～99.9%,respectively.This result indicates that the GAS method has higher segmentation accuracy and better stability in real-image digital sieving.关键词
无人机/砾石/自动筛分/级配曲线/图像处理/粒径Key words
unmanned aerial vehicle(UAV)/gravel/automatic screening/grading curves/image processing/particle size分类
水利科学引用本文复制引用
高仕赵,毛海英,范庆来,冀自青..基于无人机遥感平台的砾石自动筛分方法[J].华中科技大学学报（自然科学版）,2025,53(3):23-30,8.基金项目
山东省中青年专家奖励基金资助项目(BS2014SF016) （BS2014SF016）
山东省自然科学基金资助项目(ZR2019MEE010). （ZR2019MEE010）
