Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Обнаружение транспортных средств на изображениях загородных шоссе на основе метода Single shot multibox Detector

This chapter introduces the subject of statistical pattern recognition (SPR). It starts by considering how features are defined and emphasizes that the nearest neighbor algorithm achieves error rates comparable with those of an ideal Bayes’ classifier. The concepts of an optimal number of features, representativeness of the training data, and the need to avoid overfitting to the training data are stressed. The chapter shows that methods such as the support vector machine and artificial neural networks are subject to these same training limitations, although each has its advantages. For neural networks, the multilayer perceptron architecture and back-propagation algorithm are described. The chapter distinguishes between supervised and unsupervised learning, demonstrating the advantages of the latter and showing how methods such as clustering and principal components analysis fit into the SPR framework. The chapter also defines the receiver operating characteristic, which allows an optimum balance between false positives and false negatives to be achieved.

Statistical Pattern Recognition

Object detection in optical remote sensing images, being a fundamental but challenging problem in the field of aerial and satellite image analysis, plays an important role for a wide range of applications and is receiving significant attention in recent years. While enormous methods exist, a deep review of the literature concerning generic object detection is still lacking. This paper aims to provide a review of the recent progress in this field. Different from several previously published surveys that focus on a specific object class such as building and road, we concentrate on more generic object categories including, but are not limited to, road, building, tree, vehicle, ship, airport, urban-area. Covering about 270 publications we survey (1) template matching-based object detection methods, (2) knowledge-based object detection methods, (3) object-based image analysis (OBIA)-based object detection methods, (4) machine learning-based object detection methods, and (5) five publicly available datasets and three standard evaluation metrics. We also discuss the challenges of current studies and propose two promising research directions, namely deep learning-based feature representation and weakly supervised learning-based geospatial object detection. It is our hope that this survey will be beneficial for the researchers to have better understanding of this research field.

A survey on object detection in optical remote sensing images

In this paper, we examine the role of polarimetry in synthetic aperture radar (SAR) interferometry. We first propose a general formulation for vector wave interferometry that includes conventional scalar interferometry as a special case. Then, we show how polarimetric basis transformations can be introduced into SAR interferometry and applied to form interferograms between all possible linear combinations of polarization states. This allows us to reveal the strong polarization dependency of the interferometric coherence. We then solve the coherence optimization problem involving maximization of interferometric coherence and formulate a new coherent decomposition for polarimetric SAR interferometry that allows the separation of the effective phase centers of different scattering mechanisms. A simplified stochastic scattering model for an elevated forest canopy is introduced to demonstrate the effectiveness of the proposed algorithms. In this way, we demonstrate the importance of wave polarization for the physical interpretation of SAR interferograms. We investigate the potential of polarimetric SAR interferometry using results from the evaluation of fully polarimetric interferometric shuttle imaging radar (SIR)-C/X-SAR data collected during October 8-9, 1994, over the SE Baikal Lake Selenga delta region of Buriatia, Southeast Siberia, Russia.

Polarimetric SAR interferometry

For ultrahigh resolution synthetic aperture radar (SAR), azimuth spectrum aliasing limits the application of frequency-domain autofocus algorithms. Therefore, time-domain autofocus algorithms are often used for ultrahigh resolution SAR imaging. However, the azimuth deramping operation in current time-domain autofocus algorithms may introduce an additional azimuth-dependent phase. This phase can be regarded as a part of the phase error, which significantly reduces the estimation accuracy of the phase error. To address this issue, this article proposes a new time-domain autofocus algorithm based on azimuth scaling transformation (AST) for ultrahigh resolution SAR. In this algorithm, we first adopt the azimuth scaling operation to avoid the azimuth-dependent phase so that the estimation accuracy of error can be greatly improved. Then, for the azimuth-dependent shifts caused by the azimuth scaling operation, we adopt the alignment processing to remove them in azimuth-time domain. Finally, we can estimate the error accurately from the aligned signal. The simulation and measured data were processed to verify the effectiveness of the algorithm.

Time-Domain Autofocus for Ultrahigh Resolution SAR Based on Azimuth Scaling Transformation

For the multibaseline (MB) synthetic aperture radar (SAR) interferometry (InSAR), MB phase unwrapping (PU) is an important step. With the rapid development of MB InSAR, the size of the datasets from the MB InSAR system is becoming increasingly larger. Under the situation of "bigdata", MB PU may face new problems with insufficient computing resources, or take too much running time to get the PU result. In order to deal with such case, we propose a convex hull and cluster-analysis based fast large-scale MB PU method (CCFLS) with enlightened by the single baseline (SB) PU method (CHFLS) from H. Yu [1]. CCFLS uses the clustering phenomenon of the MB residues to generate the convex hull of residues set with balance polarity, and avoids spending the computation resources on the area within the convex hull, so that the high-precision PU solution can be quickly obtained. The theoretical analysis and experiment results indicate that CCFLS can effectively reduce memory consumption and calculation time.

A Convex Hull and Cluster-Analysis Based Fast Large-Scale Phase Unwrapping Method for Multibaseline Sar Interferograms

Wideband Radar Detection for Maneuvering Multi-Target: Wideband Radar Detection for Maneuvering Multi-Target

Synthetic Aperture Radar (SAR) imaging with a non-zero (forward) squint angle is capable of providing a longer time for reaction than that of the broadside mode. However, due to the large squint angle, there will be severe coupling between range and azimuth samples in the echoed data, which is known as the problematic Range Cell Migration (RCM) in the SAR community. Especially when the SAR sensor mounted on an airborne platform encounters unexpected motion deviations/errors, the coupling becomes more complicated, and it is difficult to differentiate the systematic RCM for the SAR Image Formation Processing (IFP) and the non-systematic RCM error to be compensated. To this end, a novel and accurate SAR imaging algorithm is proposed in this paper to facilitate the processing of airborne SAR data collected at a high-squint angle. Firstly, the proposed algorithm is established under a Fast Time-Domain Back-Projection (FTDBP) framework for the SAR IFP. FTDBP paves the way to avoid the complicated processing for the systematic RCM as for the conventional SAR IFP in the Doppler processing manner. It is capable of generating a high-resolution SAR image efficiently under more general geometries and configurations. Secondly, regarding the non-systematic RCM errors, the proposed algorithm realizes the compensation by correcting both the Non-systematic Range Cell Migration (NsRCM), as well as Azimuthal Phase Error (APE) in a coherent manner. It is consequently capable of auto-calibrating the effects of the motion error completely without being dependent on the airborne navigation unit. Finally, both simulated and raw data collected by the airborne squinted SAR are applied to evaluate the proposed algorithm. Comparisons with conventional algorithms are carried out to reveal the superiority of the proposed algorithm.

/pdf/coherent-auto-calibration-of-ape-and-nsrcm-under-fast-back-3vym0cn5uf.pdf

Coherent auto-calibration of APE and NsRCM under fast back-projection image formation for airborne SAR imaging in highly-squint angle

A new way of looking at Loffeld's bistatic formula (LBF) is presented for tandem configuration in this letter. The factors that affect the precision of the spectrum are obtained through the comparison with another analytical one. It has been proved that the cosine of the half bistatic angle plays a more important role respect to the other factor, which is whether the baseline to range ratio is equal to the tangent of the half bistatic angle or not. As long as the cosine of the half bistatic angle is very close to one, the LBF spectrum is of high quality, and it does not have a direct influence by the length of the baseline (baseline-to-range ratio) or the size of the squint angle. The factors that affect the precision of the spectrum are discussed in detail through simulated experiments.

Mengdao Xing

Papers

Time-Domain Autofocus for Ultrahigh Resolution SAR Based on Azimuth Scaling Transformation

A Convex Hull and Cluster-Analysis Based Fast Large-Scale Phase Unwrapping Method for Multibaseline Sar Interferograms

Wideband Radar Detection for Maneuvering Multi-Target: Wideband Radar Detection for Maneuvering Multi-Target

Coherent auto-calibration of APE and NsRCM under fast back-projection image formation for airborne SAR imaging in highly-squint angle

A New Look at Loffeld's Bistatic Formula in Tandem Configuration