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

Due to the independence of azimuth-invariant assumption of an echo signal, time-domain algorithms have significant performance advantages for missile-borne synthetic aperture radar (SAR) focusing with curve moving trajectory. The Cartesian factorized back projection (CFBP) algorithm is a newly proposed fast time-domain implementation which can avoid massive interpolations to improve the computational efficiency. However, it is difficult to combine effective and efficient data-driven motion compensation (MOCO) for achieving high focusing performance. In this paper, a new data-driven MOCO algorithm is developed under the CFBP framework to deal with the motion error problem for missile-borne SAR application. In the algorithm, spectrum compression is implemented after a CFBP process, and the SAR images are transformed into the spectrum-compressed domain. Then, the analytical image spectrum is obtained by utilizing wavenumber decomposition based on which the property of motion induced error is carefully investigated. With the analytical image spectrum, it is revealed that the echoes from different scattering points are aligned in the same spectrum range and the phase error becomes a spatial invariant component after spectrum compression. Based on the spectrum-compressed domain, an effective and efficient data-driven MOCO algorithm is accordingly developed for accurate error estimation and compensation. Both simulations of missile-borne SAR and raw data experiment from maneuvering highly-squint airborne SAR are provided and analyzed, which show high focusing performance of the proposed algorithm.

https://www.mdpi.com/2072-4292/13/8/1462/pdf

Processing Missile-Borne SAR Data by Using Cartesian Factorized Back Projection Algorithm Integrated with Data-Driven Motion Compensation

To achieve wide-swath and full azimuth resolution in space-borne Synthetic Aperture Radar (SAR), a space-borne radar usually utilizes low Pulse Repetition Frequency (PRF) with a small aperture in azimuth. Consequently, the azimuth ambiguities are inevitable. As one solution to this problem, Space domain filter can be utilized to remove the azimuth ambiguities based on the configuration of micro-satellites distributing along track. In this paper, an innovative method is proposed to simplify the calculation of the filter weight vectors. Simulation shows the validity of the method.

Suppression of Azimuth Ambiguities with Constellation of Micro-satellites

In airborne case, synthetic aperture radar images generally suffer from the deterioration because of the unknown phase error caused by unstable platform and atmosphere perturbation. To obtain the phase error, a novel autofocus algorithm referred to as blind homomorphic deconvolution autofocus algorithm is proposed in this study. In this method, a wavelet scaling function is used to construct a smooth subspace that is orthogonal to noise subspace. Then, the phase error can be separated and reconstructed by the generated smooth subspace based on the differences of the smoothness properties between phase error and image reflectivity. Compared with the traditional autofocus methods, the proposed method does not require an iterative estimation. Thus, the computational complexity can be significantly reduced. Simulation and real data processing results validate the effectiveness of the proposed method.

https://ietresearch.onlinelibrary.wiley.com/doi/pdf/10.1049/iet-rsn.2014.0399

Autofocus algorithm using blind homomorphic deconvolution for synthetic aperture radar imaging

In a short observation time, after the range alignment and phase adjustment, the motion of a target can be approximated as a uniform rotation. The radar observing process can be simply described as multiplying an observation matrix on the ISAR image, which can be thought of as a linear system. It is known that the longer observation time is, the higher cross-range resolution is. In order to deal with the conflict between short observation time and high cross-range resolution, we introduce Kalman filtering (KF) into the ISAR imaging and propose a novel method to reconstruct a high-resolution image with short time observed data. As KF has excellent reconstruction performance, it leads to a good application in ISAR image reconstruction. At each observation aperture, the reconstructed image denotes the state vector in KF at the aperture time. It is corrected by a two-step KF process: prediction and update. As iteration continues, the state vector is gradually corrected to a well-focused high-resolution image. Thus, the proposed method can obtain a high-resolution image in a short observation time. Both simulated and real data are applied to demonstrate the performance of the proposed method.

ISAR Signal Tracking and High-Resolution Imaging by Kalman Filtering

For a circular trace scanning synthetic aperture radar (CTSSAR) system, the conventional quadratic range approximation method becomes inefficient when the squint angle exists and the integrated aperture is long. To improve the accuracy in the approximation of the range history for CTSSAR, a modified hyperbolic range equation is introduced to derive the corresponding two-dimensional spectrum and further develop the imaging algorithm for CTSSAR. Promising results prove the effectiveness of the proposed method.

Mengdao Xing

Papers

Processing Missile-Borne SAR Data by Using Cartesian Factorized Back Projection Algorithm Integrated with Data-Driven Motion Compensation

Suppression of Azimuth Ambiguities with Constellation of Micro-satellites

Autofocus algorithm using blind homomorphic deconvolution for synthetic aperture radar imaging

ISAR Signal Tracking and High-Resolution Imaging by Kalman Filtering

Imaging algorithm for circular trace scanning synthetic aperture radar using modified hyperbolic range equation