Automatic target recognition for hyperspectral imagery using high-order statistics

Due to recent advances in hyperspectral imaging sensors many subtle unknown signal sources that cannot be resolved by multispectral sensors can be now uncovered for target detection, discrimination, and identification. Because the information about such sources is generally not available, automatic target recognition (ATR) presents a great challenge to hyperspectral image analysts. Many approaches developed for ATR are based on second-order statistics in the past years. This paper investigates ATR techniques using high order statistics. For ATR in hyperspectral imagery, most interesting targets usually occur with low probabilities and small population and they generally cannot be described by second-order statistics. Under such circumstances, using high-order statistics to perform target detection have been shown by experiments in this paper to be more effective than using second order statistics. In order to further address a challenging issue in determining the number of signal sources needed to be detected, a recently developed concept of virtual dimensionality (VD) is used to estimate this number. The experiments demonstrate that using high-order statistics-based techniques in conjunction with the VD to perform ATR are indeed very effective     

Effects of polarization and resolution on SAR ATR

Lincoln Laboratory is investigating the detection and classification of stationary ground targets using high resolution, fully polarimetric, synthetic aperture radar (SAR) imagery. A study is summarized in which data collected by the Lincoln Laboratory 33 GHz SAR were used to perform a comprehensive comparison of automatic target recognition (ATR) performance for several polarization/resolution combinations. The Lincoln Laboratory baseline ATR algorithm suite was used, and was optimized for each polarization/resolution case. Both the HH polarization alone and the optimal combination of HH, HV, and VV were evaluated; the resolutions evaluated were 1 ft/spl times/1 ft and 1 m/spl times/1 m. The data set used for this study contained approximately 74 km/sup 2/ of clutter (56 km/sup 2/ of mixed clutter plus 18 km/sup 2/ of highly cultural clutter) and 136 tactical target images (divided equally between tanks and howitzers).     

基于子空间投影的未知背景航拍高光谱图像恒虚警目标检测

针对航拍高光谱图像中未知背景地物特征条件下小目标的检测问题,给出一种检测算法。利用目标的低概率特性,通过模糊聚类获取高光谱图像中背景的光谱特性;然后将高光谱数据向背景光谱信号的正交子空间及目标信号子空间投影以抑制背景和噪声信号;最后在特征层利用广义似然比检验构造出具有恒虚警特性的检测器,完成融合检测过程。理论分析和实验结果表明了算法的有效性。     

一种针对高光谱图象识别与分类的二进编码方法

针对高光谱遥感图像中地物识别与分类问题,在对传统二进编码方法的优缺点进行分析的基础上,给出了一种新的二进编码策略,并给出了相应的快速匹配算法。该算法原理简单,易于实现,能够提高不同地物的可分性,试验结果表明,算法可以得到较好的分类效果。     

基于改进Yolov3 算法的舰船目标检测识别系统

针对复杂战场环境下对海目标检测识别的需求,设计了一种基于改进Yolov3 算法的海面舰船目标实 时检测识别系统。使用微调分类网络、增加训练尺度、聚类目标边框维度、二级特征分类等方法对Yolov3 检 测识别网络模型进行了优化,在提高识别精度的同时有效降低了漏检率和虚警率。实验结果表明,优化后的网 络模型在自建的舰船图像数据库中将检测识别平均准确率提高到了79.3%,对真实海上航拍视频中舰船目标识 别的平均准确率达到了81% 以上。     

Superresolution HRR ATR with high definition vector imaging

A new 1-D template-based automatic target recognition (ATR) algorithm is developed and tested on high range resolution (HRR) profiles formed from synthetic aperture radar (SAR) images of targets taken from the Moving and Stationary Target Acquisition and Recognition (MSTAR) data set. In this work, a superresolution technique known as High Definition Vector Imaging (HDVI) is applied to the HRR profiles before the profiles are passed through ATR classification. The new I-D ATR system using HDVI demonstrates significantly improved target recognition compared with previous I-D ATR systems that use conventional image processing techniques. This improvement in target recognition is quantified by improvement in probability of correct classification (PCC). More importantly, the application of HDVI to HRR profiles helps to maintain the same ATR performance with reduced radar resource requirements     

Spectral-spatial target detection based on data field modeling for hyperspectral data

Target detection is always an important application in hyperspectral image processing field. In this paper, a spectral-spatial target detection algorithm for hyperspectral data is proposed. The spatial feature and spectral feature were unified based on the data filed theory and extracted by weighted manifold embedding. The novelties of the proposed method lie in two aspects. One is the way in which the spatial features and spectral features were fused as a new feature based on the data field theory, and the other is that local information was introduced to describe the decision boundary and explore the discriminative features for target detection. The extracted features based on data field modeling and manifold embedding techniques were considered for a target detection task. Three standard hyperspectral datasets were considered in the analysis. The effectiveness of the proposed target detection algorithm based on data field theory was proved by the higher detection rates with lower False Alarm Rates (FARs) with respect to those achieved by conventional hyperspectral target detectors.     

A multiresolution approach to discrimination in SAR imagery

We develop and test a new algorithm for discriminating man-made objects from natural clutter in synthetic-aperture radar (SAR) imagery. This algorithm exploits characteristic variations in speckle pattern as image resolution is varied from course to fine. We model these variations as an autoregression in scale, and then use the autoregressive model to define a multiresolution log-likelihood ratio discriminant. We incorporate this discriminant into the existing Lincoln Laboratory SAR system for automatic target recognition (ATR), and test the augmented system by applying it to millimeter-wave SAR imagery having 0.3 m resolution and representing 56 square kilometers of terrain. At a probability of detection of 0.95, the addition of the multiresolution discriminant reduces the number of natural-clutter false alarms by a factor of six.     

红外极小目标检测算法研究

低信噪比检测技术是实现红外自动目标识别的基本前提,其性能指标将直接决定系统探测距离的远近,是反映红外低可观测目标识别能力至关重要的一项核心技术。自适应背景估计方法是实现这一目标的有效途径。本文在重点论述几种背景估计常用技术的基础上,提出了红外极小目标的形态滤波优化改进算法,通过理论分析和实验检测表明：该算法简化了形态变...     

航拍高光谱图像中基于投影的自动目标检测算法

针对未知环境条件下的高光谱图像目标检测问题进行了研究，提出了一种基于投影的自动目标检测算法。该算法通过构造正交投影算子预先对部分干扰物信息进行削弱，再以无监督的自动目标搜寻方法找到场景中可能的目标物，将图像数据向可能目标物所张成的子空间投影以增强目标物的信息，然后用匹配的方法完成检测。有效减弱了干扰物对目标检测的影响，缩小了目标搜索的范围。应用此算法对实验采集数据进行处理，取得了较好的结果。     

基于改进SSD的舰船目标精细化检测方法

以SSD为代表的主流深度学习方法在目标检测领域取得了显著的成绩，但由于该类方法只能以矩形框给出目标的概略位置，检测结果具有很大的背景冗余区域，特别是港口密集停泊的舰船在图像中会出现区域重叠，导致误检和漏检。针对以上问题，提出了一种具有旋转不变性的舰船目标精细化检测方法，该方法综合利用可变形卷积、可变形池化、旋转的边框回归和旋转的非极大值抑制等模块的优点，借鉴MobileNet架构对网络加速，通过学习密集区域目标的几何形变，有效预测目标的旋转角度，最终以旋转的矩形框给出目标的位置。实验结果表明，该算法可实现多类舰船目标类型区分和目标朝向判定的功能，有效地解决了实际应用中的目标精确定位定向难题，提高了自动目标识别的精确性，并满足工程应用的实时性要求。     

Performance Analysis of a Target Detection System Using Infrared Imagery

The automatic detection of targets in cluttered infrared imagery is considered. The environment for the problem is that of a "fire-and-forget" weapon, and the mission philosophy for such a weapon dictates that the weapon has to find one and only one target in the automatic detection phase. A detection system that meets this requirement is presented. The system uses techniques of image processing and pattern recognition, with the extension that ranking methods are used instead of thresholds to accommodate the requirement of finding one and only one target. A probability model of the system is developed to determine the system performance as a function of throughput and expressions derived for the probability that the object chosen by the system as the target is actually a target. In order to validate the theoretical results, the actual performance of the detection system on a database of 68 infrared images is determined and compared with the predicted performance of the system. It is shown that there is good correspondence between the empirical results and the theoretical performance.     

A new algorithm for the formation of ISAR images

A new technique for implementing the enhanced image processing (EIP) algorithm for the formation of inverse synthetic aperture radar (ISAR) images is presented. The EIP algorithm is required when, during the formation of an image, scattering centers on a target move out of range and/or Doppler resolution cells. This phenomenon is common for high resolution imagery of practical-sized targets. The method presented is based entirely on the fast Fourier transform (FFT) and therefore does not require the interpolation schemes that are prevalent in the standard EIP implementation. A brief review of the theory of radar imaging is presented to establish the notation for the work. Following the presentation of the new algorithm, a simple example is given to demonstrate the effectiveness of the new technique. In addition work is presented that demonstrates the processing required to reduce the sidelobes in imagery generated by the EIP technique     

Automatic recognition of ISAR ship images

Inverse synthetic aperture radar (ISAR) produces images of ships at sea which human operators can be trained to recognize. Because ISAR uses the ship's own varying angular motions (roll, pitch, and yaw) for cross-range resolution, the viewing aspect and cross-range scale factor are continually changing on time scales of a few seconds. This and other characteristics of ISAR imaging make the problem of automatic recognition of ISAR images quite distinct from the recognition of optical images. The nature of ISAR imaging of ships, and single-frame and multiple-frame techniques for segmentation, feature extraction, and classification are described. Results are shown which illustrate a capability for automatic recognition of ISAR ship imagery     

一种基于红外图像的目标自动识别算法

红外成像制导具有在各种复杂战术环境下自主搜索、捕获、识别和跟踪目标的能力,代表了当代红外制导技术的发展趋势。提出了一种红外图像预处理、跟踪、分类的自动目标识别算法,利用小波变换、形态学方法对红外图像进行预处理,提取不同频带的惯性不变矩作为特征量,利用神经网络进行分类识别,结果表明该算法具有很高的识别率,对于精确制导武器的目标识别研究具有一定的参考价值。     

基于红外视觉/激光雷达融合的目标识别与定位方法

近年来,基于可见光图像的目标识别在无人车感知领域得到了广泛应用.然而,可见光图像目标识别无法应用于弱光和黑暗环境.针对于此,提出了一种基于红外视觉/激光雷达融合的目标识别与定位算法.首先,通过基于颜色迁移的数据增强训练方法,提高了红外目标识别算法的泛化性能.继而,提出了一种基于激光雷达修正的单目深度估计方法,通过视觉图像与激光雷达点云的数据融合,实现了基于稠密深度图对目标位置的估计,提高了对小目标的定位能力.试验结果表明,与传统Yolov3目标识别算法相比,该算法平均识别准确率可提升5.8％;此外,相对定位算法将包含小目标在内的物体相对定位精度提高了13.4％.     

GMM-based target classification for ground surveillance Doppler radar

An automatic target recognition (ATR) algorithm, based on greedy learning of Gaussian mixture model (GMM) is developed. The GMMs were obtained for a wide range of ground surveillance radar targets such as walking person(s), tracked or wheeled vehicles, animals, and clutter. Maximum-likelihood (ML) and majority-voting decision schemes were applied to these models for target classification. The corresponding classifiers were trained and tested using distinct databases of target echoes, recorded by ground surveillance radar. ML and majority-voting classifiers obtained classification rates of 88% and 96%, correspondingly. Both classifiers outperform trained human operators.     

Automatic target recognition using enhanced resolution SAR data

Using advanced technology, a new automatic target recognition (ATR) system has been developed that provides significantly improved target recognition performance compared with ATR systems that use conventional synthetic aperture radar (SAR) image-processing techniques. This significant improvement in target recognition performance is achieved by using a new superresolution image-processing technique that enhances SAR image resolution (and image quality) prior to performing target recognition. A computationally efficient two-level implementation of a template-based classifier is used to perform target recognition. The improvement in target recognition performance achieved using superresolution image processing in this new ATR system is quantified     

A new approach to automatic target recognition

The fundamental problems of automatic target recognition (ATR) are discussed. A new approach to ATR is suggested that includes: a new method of scoring ATR performance, a new concept of artificial images, a new method called probing for extracting target signature knowledge from image experts, and suggestions for coping with the problem of insufficient test data and algorithm obsolescence     

基于视觉引导着陆的微小型四旋翼设计及实现

为了增强四旋翼的自主性,借用视觉导航方法要求在GPS丢失信号或信号精度不高的情况下进行辅助定位,使得四旋翼能够在某些特殊需求场合自动着陆。针对Harris角点检测算法计算量大的问题,设计了一种快速、准确的标志图像识别算法,将一个区域内的所有角点转换成一个单值角点,替代了Harris算法中的局部非极大值抑制算法,减小了计算量,并进一步设计了视觉模式下的姿态与位置转换算法和视觉着陆控制算法。为了验证所设计算法的可行性和有效性,基于STM32单片机搭建了微型四旋翼实验平台系统,并实时在线进行了性能实验,结果表明,标志图像识别算法能正确检测和识别出标志图像,视觉着陆算法也成功实现了四旋翼实验平台的着陆功能。