TY - JOUR
T1 - Fuzzy clustering based ET image fusion
AU - Yue, Shihong
AU - Wu, Teresa
AU - Pan, Jian
AU - Wang, Huaxiang
N1 - Funding Information:
This work is supported by the National Science Foundation of China under Grant Nos. 61774014, 60572065, 60772080 and the National Science Foundation of Tianjin under Grant No. 08JCYBJC13800.
PY - 2013
Y1 - 2013
N2 - Electrical tomography (ET) is a technique to visually reconstruct inhomogeneous medium distributions by injecting currents or voltages at the boundary of the medium and measuring the resulted changes in the investigated fields. The ET techniques have been widely used in industrial practices owing to the low cost, rapid response time, non-existent radiation exposure, and non-intrusive characteristics comparing to other tomographic modalities. However, the spatial resolution of ET images using single modality or single-driven patterns (adjacent pattern vs. opposite pattern for imaging reconstruction) is low, which may limit its applications. In this research, the application of fuzzy clustering based fusion techniques for ET imaging is studied. Both multi-modality imaging and multi-driven patterns are of interest. Specifically, two modality images are fused: Electrical Capacitance Tomography (ECT), which performs well for imaging material of large permittivity difference, and Electrical Resistance Tomography (ERT), which is suited for imaging materials having large conductivity differences. The research also explores the fusion of adjacent and opposite patterns for either ECT or ERT modalities. Experiments show that the proposed method can construct high quality ET images by discovering the strong complementary natures of the modalities and/or driven patterns.
AB - Electrical tomography (ET) is a technique to visually reconstruct inhomogeneous medium distributions by injecting currents or voltages at the boundary of the medium and measuring the resulted changes in the investigated fields. The ET techniques have been widely used in industrial practices owing to the low cost, rapid response time, non-existent radiation exposure, and non-intrusive characteristics comparing to other tomographic modalities. However, the spatial resolution of ET images using single modality or single-driven patterns (adjacent pattern vs. opposite pattern for imaging reconstruction) is low, which may limit its applications. In this research, the application of fuzzy clustering based fusion techniques for ET imaging is studied. Both multi-modality imaging and multi-driven patterns are of interest. Specifically, two modality images are fused: Electrical Capacitance Tomography (ECT), which performs well for imaging material of large permittivity difference, and Electrical Resistance Tomography (ERT), which is suited for imaging materials having large conductivity differences. The research also explores the fusion of adjacent and opposite patterns for either ECT or ERT modalities. Experiments show that the proposed method can construct high quality ET images by discovering the strong complementary natures of the modalities and/or driven patterns.
KW - Driven pattern
KW - Electrical tomography
KW - Image fusion
KW - Modality
UR - http://www.scopus.com/inward/record.url?scp=84887826525&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84887826525&partnerID=8YFLogxK
U2 - 10.1016/j.inffus.2012.09.004
DO - 10.1016/j.inffus.2012.09.004
M3 - Article
AN - SCOPUS:84887826525
VL - 14
SP - 487
EP - 497
JO - Information Fusion
JF - Information Fusion
SN - 1566-2535
IS - 4
ER -