Studied the principle of transient electromagnetic method in coalmine and solved the computation of the whole time apparent resistivity and the relation between apparent resistivity and exploration depth and so on. St...Studied the principle of transient electromagnetic method in coalmine and solved the computation of the whole time apparent resistivity and the relation between apparent resistivity and exploration depth and so on. Studied the work method of transient electromagnetic method in coalmine and obtained reasonable arrangement way. Studied data processing and explanation method of transient electromagnetic method and obtained high quality electric section. Finally the purpose to detect water-bearing body and water-bearing structure in front of roadway in advance, and detect the water-bearing property of the roof and floor rock layer of coal face were realized by use of transient electromagnetic method.展开更多
Roof vertex information is vital for 3D roof structures.Reconstructing 3D roof structures from point cloud data using traditional methods remains a challenge because their extracted roof vertices are affected by uncer...Roof vertex information is vital for 3D roof structures.Reconstructing 3D roof structures from point cloud data using traditional methods remains a challenge because their extracted roof vertices are affected by uncertainty and additional errors from roof plane segmentation and supplementary sub-steps for extracting primitives.In this study,instead of segmenting roof planes and then extracting primitives based on them,a flexible rule-based method is proposed to directly detect the vertices of building roofs from point cloud data without the requirement of training data.The point cloud data is first voxelized with a dominant direction-based rotation.Based on the different features of the interior roof points and vertices,rules for voxel filtering and structure line determination are defined to extract the roof vertices.The experimental results on a custom dataset in Trondheim,Norway demonstrate that the proposed method can effectively and accurately extract roof vertices from point cloud data.The comparative experimental results with an unfine-tuned deep learning-based method on custom and benchmark datasets with different point densities further show that the proposed method has good generalization and can adapt to changes of datasets.展开更多
文摘Studied the principle of transient electromagnetic method in coalmine and solved the computation of the whole time apparent resistivity and the relation between apparent resistivity and exploration depth and so on. Studied the work method of transient electromagnetic method in coalmine and obtained reasonable arrangement way. Studied data processing and explanation method of transient electromagnetic method and obtained high quality electric section. Finally the purpose to detect water-bearing body and water-bearing structure in front of roadway in advance, and detect the water-bearing property of the roof and floor rock layer of coal face were realized by use of transient electromagnetic method.
基金supported by NTNU Digital project(project No.81771593).
文摘Roof vertex information is vital for 3D roof structures.Reconstructing 3D roof structures from point cloud data using traditional methods remains a challenge because their extracted roof vertices are affected by uncertainty and additional errors from roof plane segmentation and supplementary sub-steps for extracting primitives.In this study,instead of segmenting roof planes and then extracting primitives based on them,a flexible rule-based method is proposed to directly detect the vertices of building roofs from point cloud data without the requirement of training data.The point cloud data is first voxelized with a dominant direction-based rotation.Based on the different features of the interior roof points and vertices,rules for voxel filtering and structure line determination are defined to extract the roof vertices.The experimental results on a custom dataset in Trondheim,Norway demonstrate that the proposed method can effectively and accurately extract roof vertices from point cloud data.The comparative experimental results with an unfine-tuned deep learning-based method on custom and benchmark datasets with different point densities further show that the proposed method has good generalization and can adapt to changes of datasets.
