煤矿地质无人机航空摄影测量图像目标自动检测方法研究
Research on Automatic Target Detection Method for Aerial Photogrammetry Images of Coal Mine Geological Unmanned Aerial Vehicles
魏跃东1
作者信息
- 1. 山西晋神沙坪煤业有限公司,山西 忻州 036500
- 折叠
摘要
图像目标检测是无人机航空摄影测量图像处理中的重要环节,其能为煤矿地质勘查决策提供依据,但测量图像中包含的小目标较多,图像目标检测难度较大,因此,设计一种煤矿地质无人机航空摄影测量图像目标的自动检测方法.该方法采用空间域灰度变换法对原始测量图像进行灰度化处理,通过均值滤波法对测量图像进行平滑滤波,根据统计测量图像直方图分割测量图像,利用神经网络算法提取测量图像的目标特征,以实现煤矿地质无人机航空摄影测量图像目标的自动检测.由实例应用分析可知,该设计方法的召回率在97%以上,错检率不超过1%,表明煤矿地质无人机航空摄影测量图像目标的检测精准度较高.
Abstract
Image target detection is an important part of unmanned aerial vehicle aerial photogrammetry image processing,which can provide a basis for coal mine geological exploration decisions.However,the measurement images contain many small targets,and image object detection is difficult.Therefore,a method for automatic detection of targets in coal mine geological unmanned aerial vehicle aerial photogrammetry images is designed.This method uses spatial domain grayscale transformation to grayscale process the original measurement image,and applies mean filtering to smooth the measurement image.The measurement image is segmented based on the histogram of the statistical measurement image,and the target features of the measurement image are extracted using neural network algorithms to achieve automatic detection of targets in coal mine geological unmanned aerial vehicle aerial photography measurement images.Through case application analysis,it can be seen that the recall rate of this design method is over 97%,and the false detection rate does not exceed 1%,indicating that the detection accuracy of coal mine geological unmanned aerial vehicle aerial photogrammetry image targets is high.
关键词
煤矿地质/无人机/航空摄影测量/图像目标/自动检测/均值滤波法/直方图Key words
coal mine geology/unmanned aerial vehicles/aerial photogrammetry/image targets/automatic detection/mean filtering method/histogram引用本文复制引用
出版年
2024
NETL
NSTL
万方数据