Detection of subpixel anomalies in multispectral infrared imageryusing an adaptive Bayesian classifier

The detection of subpixel targets with unknown spectral signatures and cluttered backgrounds in multispectral imagery is a topic of great interest for remote surveillance applications. Because no knowledge of the target is assumed, the only way to accomplish such a detection is through a search for anomalous pixels. Two approaches to this problem are examined in this paper. The first is to separate the image into a number of statistical clusters by using an extension of the well-known k-means algorithm. Each bin of resultant residual vectors is then decorrelated, and the results are thresholded to provide detection. The second approach requires the formation of a probabilistic background model by using an adaptive Bayesian classification algorithm. This allows the calculation of a probability for each pixel, with respect to the model. These probabilities are then thresholded to provide detection. Both algorithms are shown to provide significant improvement over current filtering techniques for anomaly detection in experiments using multispectral IR imagery with both simulated and actual subpixel targets     

基于光谱角背景纯化的高光谱异常检测算法

为了解决利用高光谱图像进行异常检测时结果不准确、虚警率较高的问题,提出了一种基于光谱角背景纯化的异常检测算法。该算法以局部RX算法为基础,根据光谱角距离分离出内外窗口间背景像元中的异常成分,得到纯化后的背景像元,然后进行异常检测。为验证算法的有效性,选取了两组机载可见光/红外光成像光谱仪真实高光谱数据进行仿真实验,并与经典的全局RX、局部RX算法进行对比。结果表明,与局部RX算法相比,该算法在两组数据下的曲线下面积分别提高了0.0317和0.0053。这些结果为下一步的研究方向提供了参考。     

Image interpretation and prediction in microwave diversity imaging

The microwave image of a metallic object is interpreted from a point of view based on the understanding of the interconnection between the scattering mechanisms, the data acquisition system, and the image reconstruction algorithm. From this understanding it is possible to interpret and predict microwave images reconstructed from data collected over specified and angular windows. The connection between a special scattering mechanism, edge diffraction, and its reconstructed image is established. The microwave image of an edge is two bright points whose locations correspond to the end points of the edge if the normal aspect angle is not included in the angular windows; otherwise a line joining the two end points and representing the edge will appear in the image. Experimental images of a trihedral reflector constructed from data collected over different angular windows support this approach to image interpretation and prediction     

基于协峭度张量的高光谱图像异常检测

高光谱图像中的异常像元往往具有在图像中出现的概率低和游离于背景数据云团之外的特点,如何“自动”确定这些异常像元是高光谱遥感图像处理中的一个重要研究方向。经典的高光谱异常检测方法一般从图像的统计特性入手,广泛应用的RXD异常检测算法通过计算图像的2阶统计特征,可以直接给出异常点的分布情况,算法复杂度低,但缺点是没有考虑到图像的高阶统计信息。基于独立成分分析的异常检测算法虽然考虑了高阶统计量对异常点的敏感性,但需要反复迭代提取异常成分后,再对提取后的成分进行异常检测。该文提出一种基于协峭度张量的异常检测算法,该算法不需要事先提取异常成分,可以直接对观测像元进行逐一检测,从而给出异常点的分布情况。基于模拟数据和真实数据的实验结果表明,该方法能够在检测出异常像元的同时更好地压制背景信息、减小虚警率,从而提高异常检测精度。     

Multiscale segmentation and anomaly enhancement of SAR imagery

We present efficient multiscale approaches to the segmentation of natural clutter, specifically grass and forest, and to the enhancement of anomalies in synthetic aperture radar (SAR) imagery. The methods we propose exploit the coherent nature of SAR sensors. In particular, they take advantage of the characteristic statistical differences in imagery of different terrain types, as a function of scale, due to radar speckle. We employ a class of multiscale stochastic processes that provide a powerful framework for describing random processes and fields that evolve in scale. We build models representative of each category of terrain of interest (i.e., grass and forest) and employ them in directing decisions on pixel classification, segmentation, and anomalous behaviour. The scale-autoregressive nature of our models allows extremely efficient calculation of likelihoods for different terrain classifications over windows of SAR imagery. We subsequently use these likelihoods as the basis for both image pixel classification and grass-forest boundary estimation. In addition, anomaly enhancement is possible with minimal additional computation. Specifically, the residuals produced by our models in predicting SAR imagery from coarser scale images are theoretically uncorrelated. As a result, potentially anomalous pixels and regions are enhanced and pinpointed by noting regions whose residuals display a high level of correlation throughout scale. We evaluate the performance of our techniques through testing on 0.3-m resolution SAR data gathered with Lincoln Laboratory's millimeter-wave SAR.     

基于Sobel算子的图像边缘检测研究

在图像处理中,提出了一种新的边缘检测算法,该算法首先对由Otsu算法所得图像进行形态学的去除孤立像素和闭运算操作,然后再用Sobel算子进行边缘检测,最后将所得图像与用Sobel算子直接对原始图像进行边缘检测的图像相加。最后用MATLAB对数字图像和红外图像都进行了仿真。仿真结果表明,该算法改善了单一的用Sobel算子检测对垂直与水平方向敏感,其他方向不敏感的不足,提高了Sobel边缘检测算子的性能,具有良好的检测精度。此算法不仅适用于数字图像还适用于红外图像。     

基于边缘检测和像素分类的可逆数据隐藏

为了提高嵌入算法的整体性能,提出一种基于边缘检测和像素分类的灰度图像可逆数据隐藏算法。算法按嵌入容量优先原则自适应地选择最佳阈值提取图像边缘,并根据边缘信息和指定的图像质量控制因子将像素分为平滑、弱边缘和强边缘像素3类。对平滑像素,通过增加嵌入强度的方法提高嵌入容量,预测误差直方图平移2位;对弱边缘像素,像素值最大修改量为1,从而保证图像含密质量;强边缘像素的像素值保持不变,进一步提高含密图像质量。实验结果表明,本文算法在保证图像质量的同时有效提高嵌入容量,算法的整体性能优于其它同类算法。     

一个基于视频图像检测火灾的算法

针对传统火灾检测技术在面对大空间结构建筑及复杂环境中的不足,提出一种基于视频图像检测早期火灾的算法。该算法首先对视频图像进行阈值分割,然后根据火焰的颜色特征获取其中的疑似火焰区域。在此基础上计算疑似火焰区域的圆形度,并将圆形度与其整体特性相结合进行火灾的早期探测和预报。     

像素预测误差耦合似然映射的图像伪造检测算法

为了解决当前图像伪造定位技术因使用了CFA 插值,易形成颜色插值噪声而降低分辨率,导致其难以检测微小篡改区域,使其伪造检测精度较低等不足,本文提出了像素预测误差耦合似然映射的图像伪造检测算法。首先,分析颜色滤波阵列CFA插值模型,并从图像中提取绿色分量;随后,嵌入权重因子,构造预测误差及其权重方差计算模型;根据预测误差与贝叶斯理论,定义伪造特征统计模型,识别出趋于零的特征值;最后,根据特征统计模型,建立其似然率模型,输出伪造映射,完成检测。仿真结果表明：与当前图像伪造定位机制相比,本文算法拥有更强的鲁棒性,能识别定位出微小伪造像素;且拥有更高的AUC值与理想的ROC曲线。     

Image-Based Traffic Monitoring With Shadow Suppression

For a vision-based traffic monitoring and enforcement system, shadows of moving objects often cause serious errors in image analysis due to misclassification of shadows and moving vehicles. An effective shadow suppression method is thus required to improve the accuracy of image analysis and this paper proposes a novel color-space ratio model for detecting shadow pixels in traffic imagery. The proposed approach does not require many image sequences for constructing the model. Instead the model can be easily built up using a shadow region in a single image frame. To increase the accuracy of shadow detection, we design two types of spatial analysis to verify actual shadow pixels. Comparative results show that the proposed method works better than several well-known methods. The proposed methods have been applied to an image-based traffic monitoring system for detecting shadow pixels in traffic imagery. The experimental results not only validate the feasibility of the proposed algorithm but also successfully estimate traffic parameters such as traffic flows, traffic densities, vehicle turn ratios and vehicle speeds, all with satisfactory accuracy     

Stochastic model-based processing for detection of small targets innon-Gaussian natural imagery

Stochastic background models incorporating spatial correlations can be used to enhance the detection of targets in natural terrain imagery, but are generally difficult to apply when the statistics are non-Gaussian. Chapple and Bertilone (see Opt. Commun., vol.150, p.71-76, 1998) proposed a simple stochastic model for images of natural backgrounds based on the pointwise nonlinear transformation of Gaussian random fields, and demonstrated its effectiveness and computational efficiency in modeling the textures found in natural terrain imagery acquired from airborne IR sensors. In this paper, we show how this model can be used to design algorithms that detect small targets (up to several pixels in size) in natural imagery by identifying anomalous regions of the image where the statistics differ significantly from the rest of the background. All of the model-based algorithms described here involve nonlinear spatial processing prior to the final decision threshold. Monte Carlo testing reveals that the model-based algorithms generally perform better than both the adaptive threshold filter and the generalized matched filter for detecting low-contrast targets, despite the fact that they do not require the target statistics needed for constructing the matched filter.     

基于支持向量机的飞机图像识别算法

研究了一种基于支持向量机的飞机图像识别算法。采用基于神经网络的图像边缘检测方法,该方法首先基于邻域灰度极值提取边界候选图像,然后以边界候选象素及其邻域象素的二值模式作为样本集,输入边缘检测神经网络进行训练。提取具有RST不变性的轮廓特征向量,输入支持向量机进行训练和识别。将支持向量机与传统的人工神经网络的算法进行了对比实验,实验表明基于支持向量机的飞机图像识别算法具有更好的性能。     

结合光谱相似性与相位一致模型的高分辨率遥感图像分割方法

提出一种结合高分辨率遥感图像的光谱相似性与相位一致边缘检测模型的分水岭分割方法.分水岭分割算法的性能依赖干图像边缘梯度图.利用同类地物的光谱相似性特点,可有效抑制相位一致模型边缘检测中产生的伪边缘和噪声信息,从而获得更好的遥感图像分割结果.首先基于目标像元与其邻域像元之间的光谱曲线距离之和定义光谱相似性模型,结合相位一致模型获得边缘响应强度;然后利用自动标记分水岭变换方法实现高分辨率遥感图像分割.基于此方法和其它方法进行了分割实验,并利用基于多光谱信息熵方法对分割结果进行了非监督评价和比较,同时对比分析了计算耗时.结果表明此方法可有效抑制遥感图像过分割现象,并取得较好的分割结果.     

一种基于极值的自适应中值滤波算法

图像平滑处理中,如何在去除噪声的同时完整地保留图像边缘细节一直是非线性滤波算法研究的热点问题。提出了一种基于极值的自适应中值滤波算法,该算法根据图像中某点是否为邻域极值点将全部像素分为可疑噪声与信号两类。对可疑噪声点采用包括八个一维窗口和一个二维窗口在内的不同尺度和不同方向的九个子窗口,按照各个子窗口的均方差大小,自适应选择窗口进行中值滤波;对信号点不加处理,灰度值不变。测试结果表明,该算法的滤噪特性和细节保护能力优于多级中值滤波;执行速度较快,优于经典中值滤波。     

结合光谱相似性与相位一致模型的高分辨率遥感图像分割方法

提出一种结合高分辨率遥感图像的光谱相似性与相位一致边缘检测模型的分水岭分割方法。分水岭分割算法的性能依赖于图像边缘梯度图。利用同类地物的光谱相似性特点,可有效抑制相位一致模型边缘检测中产生的伪边缘和噪声信息,从而获得更好的遥感图像分割结果。本文首先基于目标像元与其邻域像元之间的光谱曲线距离之和定义光谱相似性模型,结合相位一致模型获得边缘响应强度;然后利用自动标记分水岭变换方法实现高分辨率遥感图像分割。使用本文方法和其它方法对遥感图像进行了分割实验,并利用基于多光谱信息熵方法对分割结果进行了非监督评价和比较,同时对计算耗时进行了对比分析。结果表明本文所提方法可以有效地抑制遥感图像的过分割现象,并取得较好的分割结果。     

基于预测方式选择的可逆信息隐藏

为了提高信息隐藏算法的性能,提出一种基于预测 方式选择和直方图平移的灰度图像可逆信息隐藏算法。通过两种相对准确的预测方法对像素 进行预 测,对突变像素进行标记,标记像素不做任何处理。最后选择其中一种相对准确 的预测值方法作为最终预测方法,求出预测误差并利用预测误差扩展和直方图平 移技术进行数据嵌入。算法自适应地选择最佳预测方法,充分利用图像冗余,有 效提高嵌入容量(EC)。实验结果显示,本文算法充分利用邻近像素间的关联性 ,在保证EC的同时有效提高了峰值信噪比(PSNR)和算法的整体性能。     

Contextual Pattern Recognition Applied to Cloud Detection and Identification

The use of contextuial information in classification of satellite imagery of clouds is considered. A general contextual classification rule is developed which is then simplified under some realistic assumptions made. for the label relationships of neighboring cloud image pixels. The resulting decision rule is defined in terms of transition probabilities between pairs of neighboring pixels. Experimental evaluation of the rule confirms its superiority over the conventional Bayes rule and provides justification for the assumption made.     

Detection and analysis of change in remotely sensed imagery withapplication to wide area surveillance

A new approach to wide area surveillance is described that is based on the detection and analysis of changes across two or more images over time. Methods for modeling and detecting general patterns of change associated with construction and other kinds of activities that can be observed in remotely sensed imagery are presented. They include a new nonlinear prediction technique for measuring changes between images and temporal segmentation and filtering techniques for analyzing patterns of change over time. These methods are applied to the problem of detecting facility construction using Landsat Thematic Mapper imagery. Full scene results show the methods to be capable of detecting specific patterns of change with very few false alarms. Under all conditions explored, as the number of images used increases, the number of false alarms decreases dramatically without affecting the detection performance. It is argued that the processing gain that results in using more than two images justifies the increased computational complexity and storage requirements of our approach over single image object detection and conventional change detection techniques.     

基于模糊中值的IRFPA自适应盲元检测与补偿

红外焦平面阵列(IRFPA)的盲元既包括因材料与制造工艺的缺陷而导致的固定盲元,也包括因环境温度的漂移而出现的随机盲元.基于场景的盲元检测与补偿算法是去除这两种盲元,提高IRFPA 成像质量的有效手段.针对目前滤波类场景检测算法无法有效区分弱小点目标和随机盲元的缺陷,重点研究了随机盲元的响应特性和噪声特性,并提出了一种基于模糊中值与时域累积的盲元自适应检测与补偿算法.首先利用模糊中值滤波器从场景中提取出潜在的盲元,并通过多帧累积确定固定盲元和随机盲元的正确分布,最后对盲元进行实时补偿.实验结果证明:该算法可以有效地实现对盲元的校正,同时避免对弱小点目标的误判别.     

An efficient macroblock-based diverse and flexible prediction modes selection for hyperspectral images coding

In this paper, an efficient macroblock-based diverse and flexible prediction modes selection algorithm is proposed for coding hyperspectral images, which is inspired by the prediction scheme of H264/AVC. Here, different modes are specified for the corresponding macroblocks (16×16 pixel regions of a band) of hyperspectral images other than the whole band image using only one reference band image for prediction. Only the 4×4 mode is employed for the intra-band prediction in view of the fact that correlation coefficients of pixels separated by not more than four pixels in the spatial domain are greater than 0.65 at most cases. The optimal reference band is determined by the fast reference band selection algorithm; thereafter, the best partition of the candidate macroblock in the optimal reference band is further selected for inter-band prediction of the current macroblock. Thus, the stronger correlation in the spectral direction or in the spatial domain is utilized for the prediction of the given macroblock. With a comparably low memory requirement, the prediction coding scheme is proposed to speed up the implemental process using the fast reference band selection algorithm, the integer DCT and the quantization, which just needs the multiplication and bit-shifts operations. Several AVIRIS images are used to evaluate the performance of the algorithm. The proposed scheme outperforms the state-of-the-art 3D-based compression algorithms at lower rates. Moreover, compared with the method by using all the prediction modes of H.264/AVC, about 80% encoding time can be saved by our method under the same experimental condition.