Abstract:The erosion of soluble rock and transformation of groundwater result in the high geometrical irregularity of cavities in nature. The purpose of this paper is to quantitatively describe the shape of irregular cavities using Fourier transform analysis. Firstly, multi-level characterization indexes of the 2D section of cavities have been proposed, considering the high shape irregularity of cavities. The first level indexes includes Shape Factor (SF), Aspect Ratio (AR) and Sphere Like (SL). The second level indexes consist of Angularity Factor (AF) and Convexity Factor (CF). The third level indexes is Roundness Factor (RF). Then, the geometric contour of cavity was transformed into a 2D waveform, with an equal-angle points sampling of the cavity boundary. The geometric boundary of irregular cavity was segmented and transformed from discrete time domain to frequency domain based on discrete Fourier transform (DFT). Furthermore, we proposed Fourier shape descriptor, which was found to control the cavity shape. The relationship between Fourier shape descriptors in different sequences and multi-level characterization indexes was determined using designed and irregular nature cavities as sample data. The results demonstrated that the Fourier shape descriptor <math>D2</math>, <math>D3</math> and <math>D8</math> mainly controls multi-level characterization indexes including the overall shape, the irregularity and the roughness of the cavity boundary, respectively. Finally, the 3D characterization method of cavity shape was further proposed based on averaging the 2D sections, which hence offers a possible way to further quantitatively study the mechanical properties of surrounding rockmass effected by the irregular shape of cavities.

A Fourier Shape Descriptor Based on Multi-Level Chord Length Function

Enhanced Micro-structure Descriptor based Image Retrieval.

Mean Descriptor: a Compact Representation for 3D Model's Multiple Features

A Novel Multi-scale Fourier Descriptor Based on Plane Orthogonal Vectors for Fast Shape Retrieval

A Fourier Descriptor of 2D Shapes Based on Multiscale Centroid Contour Distances Used in Object Recognition in Remote Sensing Images

AFSMWD: A Descriptor Flexibly Encoding Multiscale and Oriented Shape Features

Full and Partial Shape Similarity Through Sparse Descriptor Reconstruction

Modified Fourier Descriptors for Shape Representation - A Practical Approach

Radius-normal Histogram and Hybrid Strategy for 3D Shape Retrieval

3D Shape Retrieval Based on Laplace Operator and Joint Bayesian Model

Data Driven Composite Shape Descriptor Design for Shape Retrieval with a VoR-Tree

Local Deep Feature Learning Framework for 3D Shape.

Thickness Histogram and Statistical Harmonic Representation for 3D Model Retrieval

A New Shape Descriptor For Object Recognition And Retrieval

Made: A Composite Visual-Based 3d Shape Descriptor

Robust Shape Retrieval Through a Novel Statistical Descriptor

Four Side Distance: A New Fourier Shape Signature

A novel MDC framework based on orientation SFQ algorithm

A new 3D shape descriptor based on rotation

Shape characterization methods of irregular cavity using Fourier analysis in tunnel

A 3D Shape Descriptor Based on Spectral Analysis of Medial Axis