Physical reality of radar targets

Huynen, J. Richard

doi:10.1117/12.140636

Cited by 23 publications

(11 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Huynen reveals the elements in the vector have some physical meanings. For detailed information, refer to [38]. Symbol Ω denotes the set of PolSAR image pixels, i.e.,…”

Section: Proposed Methodsmentioning

confidence: 99%

Auto Encoder Feature Learning with Utilization of Local Spatial Information and Data Distribution for Classification of PolSAR Image

et al. 2019

View full text Add to dashboard Cite

The distribution of data plays a key role in the designing of a machine learning model. Therefore, this paper proposes a novel auto encoder network based on the distribution of polarimetric synthetic aperture radar (PolSAR) data matrix. Designed specifically for PolSAR data matrix, the proposed mixture auto encoder (MAE) feature learning method defines data error term in the loss function according to the data distribution. Instead of the pixel itself, all pixels in the neighborhood are used as input to train the proposed MAE. Then, a corresponding classification network is also given by discarding the decoder process of the proposed MAE and connecting with a Softmax classifier. The MAE is trained using the unlabeled data, while the training process of the classification network is completed with the help of a small number of labeled pixels. In view of the phenomenon of misclassification in the predicted result image, two post-processing steps acting on local spatial are also given, which accomplished by the proposed two filters. Extensive experiments by four methods were made over three real PolSAR images including the proposed classification network. The experimental results show that introducing data distribution into the auto encoder network leads to an average 4% improvement in overall accuracy for three PolSAR images. Moreover, the post-processing steps with the proposed filters bring a new level of discrimination on the classification performance of PolSAR images.

show abstract

“…Huynen reveals the elements in the vector have some physical meanings. For detailed information, refer to [38]. Symbol Ω denotes the set of PolSAR image pixels, i.e.,…”

Section: Proposed Methodsmentioning

confidence: 99%

Auto Encoder Feature Learning with Utilization of Local Spatial Information and Data Distribution for Classification of PolSAR Image

et al. 2019

View full text Add to dashboard Cite

show abstract

“…This is the parameter C adopted in S4R, i.e., S4R also signifies a special form of GG4U. Hence, the essential difference between S4R and G4U just lies in the different definition of parameter C in ( 44) and (46). The unitary transformation is just to enable the T 13 entry contained in C G4U and finally in P S and P D .…”

Section: B Generalized Decompositionmentioning

confidence: 99%

“…The Yamaguchi four-component decomposition (Y4O) then leaves only three DoF unaccounted for: T 13 and the real part of T 23 (Re{T 23 }) entry of target coherency matrix [T ]. The same target will present differently by a simple rotation about the line of sight of radar [24], [46]. The deorientation should be first conducted on [T ] to eliminate the influences [47].…”

Section: Introductionmentioning

confidence: 99%

A Mathematical Extension to the General Four-Component Scattering Power Decomposition With Unitary Transformation of Coherency Matrix

Liu

Zhang

Liang

2020

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

As an improvement of the four-component scattering power decomposition with rotation of coherency matrix (Y4R) and extension of volume model (S4R), the general four-component decomposition with unitary transformation (G4U) was devised to make the full use of the polarimetric information in coherency matrix. This article enables an extension to G4U by deriving the scattering balance equation system in G4U to investigate the role of unitary transformation first. Despite self-contained, the scattering balance equation system in Y4R and S4R is independent of the T 13 entry of coherency matrix. To include T 13 in decomposition, the unitary transformation in G4U adds a T 13 -related but redundant balance equation into the original system. As a result, T 13 is accounted for by G4U, and we attain no exact solution to the equation system but some approximate ones. By deducing the general expression of the approximate solutions, a generalized G4U (GG4U) is then created and denoted as G(ψ). The decomposition constant ψ determines a GG4U by producing a ψ-rotated double-bounce scattering matrix. We treat this as the scattering preference of G(ψ) to characterize the physical mechanism. By assigning appropriate values to ψ, we attain GG4U of different preferences, while G(0) and G(+π/8) just correspond to S4R and G4U. A dual G4U G(−π/8) is also achieved. The duality G(±π/8)provides us an adaptive improvement to both G4U and S4R by strengthening the double-bounce scattering over urban and building area while enhancing the surface scattering over water and land area. Both theoretical derivation and experiments on ten polarimetric synthetic aperture radar data sets validate the outperformance. Nonetheless, for whatever unitary transformation employed, there is, forever, a T 13 -related residual component in GG4U. Thus, the incorporation of unitary transformation into Y4R and S4R for the full modeling of polarimetric information is impossible in theory only when the canonical scattering model with nonzero (1, 3) entry of coherency matrix is used to add the balance equation system an independent T 13 -related equation rather than a redundant one.

show abstract

“…Chen et al [35] have not evaluated the impact of target orientation on their classification. The same scatterer can be presented differently by a simple rotation about the line of sight (LOS) [37], [38]. As the result, a building may be identified as forest because orientation increases the cross-polarized scattering [39]- [41].…”

Section: Sufficiency and Deficiency Of Chen Classificationmentioning

confidence: 99%

Adaptive Model-Based Classification of PolSAR Data

Liu

Zhang

2018

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

An adaptive classification is developed as a hybrid of the eigenvector-based and the model-based target decompositions for polarimetric synthetic aperture radar (PolSAR) data. The classification adopts the canonical scattering models that widely used in model-based decompositions to provide an improvement for the well-known H/α classification. First, a correspondence principle is adopted to adaptively identify the matched canonical models. The selected models are parallelly combined based on the scattering similarity for a fine depiction of the scattering mechanism then. Twelve classes are finally obtained, and each one carries a unique symbol to show a specific scattering. The classification does not depend on a particular data set, avoids the hard partitioning, and solves the obscures in H/α. Comparison on the real PolSAR data sets with H/α and the existing scattering similarity-based classification validates the better discrimination.

show abstract

Physical reality of radar targets

Cited by 23 publications

References 0 publications

Auto Encoder Feature Learning with Utilization of Local Spatial Information and Data Distribution for Classification of PolSAR Image

Auto Encoder Feature Learning with Utilization of Local Spatial Information and Data Distribution for Classification of PolSAR Image

A Mathematical Extension to the General Four-Component Scattering Power Decomposition With Unitary Transformation of Coherency Matrix

Adaptive Model-Based Classification of PolSAR Data

Contact Info

Product

Resources

About