Decision fusion of classifiers for multifrequency PolSAR and optical data classification

Kasapoglu, N.G.; Eltoft, T.

doi:10.1109/rast.2013.6581242

Cited by 3 publications

(4 citation statements)

References 9 publications

(7 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Cui et al [63] applied decision fusion to texture features derived from polarimetric data to evaluate levees. Furthermore, in [64], an ML classifier was utilized for the first labeling with MV and qualified majority voting (QMV) as a consensual rule for fusion. Abdikan et al [65] tested four classifiers, namely SVM, RF, K-nearest neighbor and ML, for the enhancement of land use classification.…”

Section: Sar and Optical Datamentioning

confidence: 99%

Decision Fusion at Pixel Level of Multi-Band Data for Land Cover Classification—A Review

Papadopoulos,

Koukiou,

Anastassopoulos

2024

J. Imaging

View full text Add to dashboard Cite

According to existing signatures for various kinds of land cover coming from different spectral bands, i.e., optical, thermal infrared and PolSAR, it is possible to infer about the land cover type having a single decision from each of the spectral bands. Fusing these decisions, it is possible to radically improve the reliability of the decision regarding each pixel, taking into consideration the correlation of the individual decisions of the specific pixel as well as additional information transferred from the pixels’ neighborhood. Different remotely sensed data contribute their own information regarding the characteristics of the materials lying in each separate pixel. Hyperspectral and multispectral images give analytic information regarding the reflectance of each pixel in a very detailed manner. Thermal infrared images give valuable information regarding the temperature of the surface covered by each pixel, which is very important for recording thermal locations in urban regions. Finally, SAR data provide structural and electrical characteristics of each pixel. Combining information from some of these sources further improves the capability for reliable categorization of each pixel. The necessary mathematical background regarding pixel-based classification and decision fusion methods is analytically presented.

show abstract

Section: Sar and Optical Datamentioning

confidence: 99%

Decision Fusion at Pixel Level of Multi-Band Data for Land Cover Classification—A Review

Papadopoulos,

Koukiou,

Anastassopoulos

2024

J. Imaging

View full text Add to dashboard Cite

show abstract

“…The MI value is calculated with every movement. The parameters at the maximum MI value are assumed as the optimum offset, p and q, in Equation (11). In this article, the offsets are estimated using the down-sampled Pan image as the reference image and the geocoded SPAN image as the sensed image.…”

Section: Image Registrationmentioning

confidence: 99%

“…Due to the significant radiometric and geometric differences between optical and SAR data, it is practically impossible to directly compare the two data. The previous studies using both data have been focused on well integrating different types of information derived from each data for land cover and land use mapping [8][9][10][11]. The research comparing the two data, particularly the pre-disaster optical data and the post-disaster SAR data, for damage mapping has been relatively less conducted [7,12].To overcome the differences and to enable direct comparison between them, simulation methods have been suggested [7,12].…”

mentioning

confidence: 99%

Comparison of Pre-Event VHR Optical Data and Post-Event PolSAR Data to Investigate Damage Caused by the 2011 Japan Tsunami in Built-Up Areas

2018

View full text Add to dashboard Cite

Combining pre-disaster optical and post-disaster synthetic aperture radar (SAR) satellite data is essential for the timely damage investigation because the availability of data in a disaster area is usually limited. This article proposes a novel method to assess damage in urban areas by analyzing combined pre-disaster very high resolution (VHR) optical data and post-disaster polarimetric SAR (PolSAR) data, which has rarely been used in previous research because the two data have extremely different characteristics. To overcome these differences and effectively compare VHR optical data and PolSAR data, a technique to simulate polarization orientation angles (POAs) in built-up areas was developed using building orientations extracted from VHR optical data. The POA is an intrinsic parameter of PolSAR data and has a physical relationship with building orientation. A damage level indicator was also proposed, based on the consideration of diminished homogeneity of POA values by damaged buildings. The indicator is the difference between directional dispersions of the pre and post-disaster POA values. Damage assessment in urban areas was conducted by using the indicator calculated with the simulated pre-disaster POAs from VHR optical data and the derived post-disaster PolSAR POAs. The proposed method was validated on the case study of the 2011 tsunami in Japan using pre-disaster KOMPSAT-2 data and post-disaster ALOS/PALSAR-1 data. The experimental results demonstrated that the proposed method accurately simulated the POAs with a root mean square error (RMSE) value of 2.761 • and successfully measured the level of damage in built-up areas. The proposed method can facilitate efficient and fast damage assessment in built-up areas by comparing pre-disaster VHR optical data and post-disaster PolSAR data.superior to an optical sensor due to its relative insensitivity to the weather and illumination conditions. Several studies have investigated disaster-induced damage using pre and post-disaster SAR data [1][2][3][4][5][6]. Recently, damage investigation has been improved using fully polarimetric SAR (PolSAR) data because it allows more efficient results with its additional polarization scattering mechanism compared to a single polarization mode [5,6].However, pre-disaster SAR data is less abundant than pre-disaster optical data because optical sensors have traditionally been used. Therefore, the combination of pre-disaster optical data and post-disaster SAR data is regarded as desirable for rapid damage assessment [7]. Due to the significant radiometric and geometric differences between optical and SAR data, it is practically impossible to directly compare the two data. The previous studies using both data have been focused on well integrating different types of information derived from each data for land cover and land use mapping [8][9][10][11]. The research comparing the two data, particularly the pre-disaster optical data and the post-disaster SAR data, for damage mapping has been relatively less conducted [7,12].To overco...

show abstract

“…Each target feature is assigned to each classifier, which votes the recognized target identity (ID). Targets with the maximal votes are finally recognized [52][53][54]. Logical AND/OR operation after Bayesian classification was proposed to fuse SAR-EO images for earth surface classification [55].…”

Section: Introductionmentioning

confidence: 99%

Double Weight-Based SAR and Infrared Sensor Fusion for Automatic Ground Target Recognition with Deep Learning

2018

View full text Add to dashboard Cite

This paper presents a novel double weight-based synthetic aperture radar (SAR) and infrared (IR) sensor fusion method (DW-SIF) for automatic ground target recognition (ATR). IR-based ATR can provide accurate recognition because of its high image resolution but it is affected by the weather conditions. On the other hand, SAR-based ATR shows a low recognition rate due to the noisy low resolution but can provide consistent performance regardless of the weather conditions. The fusion of an active sensor (SAR) and a passive sensor (IR) can lead to upgraded performance. This paper proposes a doubly weighted neural network fusion scheme at the decision level. The first weight (α) can measure the offline sensor confidence per target category based on the classification rate for an evaluation set. The second weight (β) can measure the online sensor reliability based on the score distribution for a test target image. The LeNet architecture-based deep convolution network (14 layers) is used as an individual classifier. Doubly weighted sensor scores are fused by two types of fusion schemes, such as the sum-based linear fusion scheme (αβ-sum) and neural network-based nonlinear fusion scheme (αβ-NN). The experimental results confirmed the proposed linear fusion method (αβ-sum) to have the best performance among the linear fusion schemes available (SAR-CNN, IR-CNN, α-sum, β-sum, αβ-sum, and Bayesian fusion). In addition, the proposed nonlinear fusion method (αβ-NN) showed superior target recognition performance to linear fusion on the OKTAL-SE-based synthetic database.Remote Sens. 2018, 10, 72 2 of 24 by considering both feature extractors and classifiers to cope with IR variations. The Markov tree feature [13], IR wavelet feature [14], scale invariant feature transform (SIFT) [15], histogram of oriented gradients (HOG) [16], and moment features [17,18] are recently proposed infrared features that show promising recognition results on their own applications. Simple machine learning-based classification methods, such as the nearest neighbor classifier [15], Bayesian classifier, conventional neural network, Adaboost [19], and support vector machine (SVM) [16] are used frequently to discriminate the target features for classification.SAR can measure the electromagnetic scattering property of targets under any weather and light conditions [20]. This method is used frequently to recognize a range of targets because it provides a strong radar cross section (RCS) and shape information of non-stealth targets. On the other hand, it produces many false recognitions due to speckle noise [21]. Various SAR features, such as polarimetry and transformations (log-polar, Fourier, and wavelet), are used to discriminate SAR targets [20,[22][23][24]. The standard deviation, fractal dimension, and weighted-rank fill ratio are the basic SAR features proposed in the Lincoln laboratory [22]. A genetic algorithm-based SAR feature selection method has been used to select the optimal features for target recognition [25]. Targets can be recognized by clas...

show abstract

Decision fusion of classifiers for multifrequency PolSAR and optical data classification

Cited by 3 publications

References 9 publications

Decision Fusion at Pixel Level of Multi-Band Data for Land Cover Classification—A Review

Decision Fusion at Pixel Level of Multi-Band Data for Land Cover Classification—A Review

Comparison of Pre-Event VHR Optical Data and Post-Event PolSAR Data to Investigate Damage Caused by the 2011 Japan Tsunami in Built-Up Areas

Double Weight-Based SAR and Infrared Sensor Fusion for Automatic Ground Target Recognition with Deep Learning

Contact Info

Product

Resources

About