Three-dimensional Reconstruction Method Of Coal Based On Stereo Vision Research

With the rapid development of Industry 4.0 and “Made in China 2025”strategy,modern port replaced traditional transportation scheme gradually,wisdom port has became the focus of national construction goals.Under the trend of the increasingly prefect of intelligent,information and automatic technology,the three-dimensional reconstruction model is becoming more and more important in the social appeal and enterprise production.Three-dimensional modeling of coal yard is an important link in the construction of intelligent port,Which determines how to develop automatic stacking and reclaiming strategies.It has important guiding significance for intelligent stacking technology.This paper takes the coal yard as studied background,the multi view stereo 3D reconstruction as theoretical principle,the UAV as Image data acquisition platform.This paper deeply analysed each link of the theory of multi view 3D reconstruction,then proposed 3D reconstruction of coal yard method based on stereo visual.Utilized multi-angle image obtained by UAV equipped with camera,and combined with the auxiliary parameter information in the image EXIF to filter and cluster for view.How to extract the stable feature points in the complex scene image,and how to get the local information of the image optimally when the image has scale transformation and fuzzy transformation.Especially for the coal pile image,its color is single,and the detection of information is missing serious.Aiming at the difficult problem of extracting characteristic points of coal heap,it is always a prerequisite to obtain as much feature information as possible for 3D reconstruction;This paper,on the basis of a huge amount of literature,combined with the geometrical feature of coal heap and carried out several experiments,proposed the feature matching algorithm which is based on adaptive scale Harris-Laplace feature point detection from UVA image auxiliary information and SIFT feature description.This algorithm can detect and match the feature of coal heap image,which increased the number of feature points and point cloud of reconstructed 3D model,and then ensure the integrity and accuracy of the model.In the stage of motion recovery,this paper has intensive study on incremental bundler spare reconstruction method and batch reconstruction method.In this paper,by using auxiliary information such as GPS in UAV image,this paper processed the image in blocking-clustering way,and combined the image projection to calculate image overlap degree between images.For inefficient and unstable drawbacks of large-scale image incremental 3D reconstruction algorithm,so the 3D reconstruction of the coal yard used batch reconstruction algorithm based on undirected graph trifocal matching.Finally,the result obtained sparse 3D point cloud of coal yard and camera parameters,and guaranteed the high efficiency of processing speed,reliability of algorithm and accuracy of data.In the end,using the algorithm of a port yard for 3-dimensional reconstruction,the obtained 3D geometric information meets the current requirements.The experimental results show that the proposed algorithm is feasible and practical,and has great research significance and application value.