UAV Image Acquisition And Information Extraction Applied Research

In order to fully rectify and standardize the development of mineral resources, curbillegal mining act, regulate mineral resources more effectively, the use of advancedtechnology means the exploitation of mineral resources and environmentalmanagement has become essential. As satellite reception of data has some limitations,UAV remote sensing technology can make up for the inadequacies of large-scaleaerial photography and satellite remote sensing measurements, significantly enhancethe monitoring efforts of mineral resources, the traditional single classificationmethod not only lead to lower the high resolution remote sensing image classificationaccuracy, but also result in a large number of spatial data redundancy and wasting ofresources.This paper studies the technology of the mineral resources information extractionbased on the UAV remote sensing. First select the Rare Earth Mine in Jiangxiprovince Ganzhou city Longnan county as the study area, describes the compositionof UAV remote sensing systems, aerial processes and the method of data processing.Through route designing, implementation flying, image preprocessing, imagematching and empty three calculating etc., to obtain the UAV remote sensinghigh-resolution image of the study area, finally get the surveyed area orthophotosthrough DEM generation.This paper lists three remote sensing image segmentation method, for miningspecial feature information, gets the best segmentation scale. Compare the results ofdifferent segmentation scales. Introduce the theory of multi-scale segmentation.Through the object-oriented multi-scale segmentation technique,creating differentclassification levels, and then sampled all kinds of surface features on ahigh-resolution imagery, access to spectral information and spatial structureinformation of surface features, select the feature information of target objects such as woodland, bare land, mine buildings, tailings and isolated pools, by the method ofcombining multiple features, define classification rules. Finally classify automatic bythe method of nearest neighbor method. Evaluate the accuracy of the classificationresults.Finally concludes the mining of mineral resources covering informationextraction technology workflow based on UAV remote sensing. Studies a number ofkey and difficult problems in information extraction process, Summarizes theclassification rules of the mine covers information automatically extraction.