Automatic target recognition and tracking in forward-looking infrared image sequences with a complex background

This paper presents a technique for automatic airborne target recognition and tracking in forward-looking infrared (FLIR) images with a complex background. An image splitting and merging method is applied for detecting target signals. The presence of a complex background due to clouds and sun glint generates clutter in the image with the resulting possibility of false alarms. A Bayesian classifier trained using the NMI (normalized moment of inertia) feature is proposed for efficient clutter rejection. After classification, target candidates are entered into a tracking filter. As an efficient and robust multi-target tracking filter in cluttered environments, the JDC-JIHPDAF is proposed. Experimental results using a wide range of real FLIR images ensure reliable classification and automatic target recognition performance.     

Object detection using image reconstruction with PCA

In this paper, we present an object detection system and its application to pedestrian detection in still images, without assuming any a priori knowledge about the image. The system works as follows: in a first stage a classifier examines each location in the image at different scales. Then in a second stage the system tries to eliminate false detections based on heuristics. The classifier is based on the idea that Principal Component Analysis (PCA) can compress optimally only the kind of images that were used to compute the principal components (PCs), and that any other kind of images will not be compressed well using a few components. Thus the classifier performs separately the PCA from the positive examples and from the negative examples; when it needs to classify a new pattern it projects it into both sets of PCs and compares the reconstructions, assigning the example to the class with the smallest reconstruction error. The system is able to detect frontal and rear views of pedestrians, and usually can also detect side views of pedestrians despite not being trained for this task. Comparisons with other pedestrian detection systems show that our system has better performance in positive detection and in false detection rate. Additionally, we show that the performance of the system can be further improved by combining the classifier based on PCA reconstruction with a conventional classifier using a Support Vector Machine.     

Robust text detection in natural scenes using text geometry and visual appearance

This paper proposes a new two-phase approach to robust text detection by integrating the visual appearance and the geometric reasoning rules. In the first phase, geometric rules are used to achieve a higher recall rate. Specifically, a robust stroke width transform (RSWT) feature is proposed to better recover the stroke width by additionally considering the cross of two strokes and the continuousness of the letter border. In the second phase, a classification scheme based on visual appearance features is used to reject the false alarms while keeping the recall rate. To learn a better classifier from multiple visual appearance features, a novel classification method called double soft multiple kernel learning (DS-MKL) is proposed. DS-MKL is motivated by a novel kernel margin perspective for multiple kernel learning and can effectively suppress the influence of noisy base kernels. Comprehensive experiments on the benchmark ICDAR2005 competition dataset demonstrate the effectiveness of the proposed two-phase text detection approach over the state-of-the-art approaches by a performance gain up to 4.4% in terms of F-measure.     

Monitoring activities from multiple video streams: establishing acommon coordinate frame

Monitoring of large sites requires coordination between multiple cameras, which in turn requires methods for relating events between distributed cameras. This paper tackles the problem of automatic external calibration of multiple cameras in an extended scene, that is, full recovery of their 3D relative positions and orientations. Because the cameras are placed far apart, brightness or proximity constraints cannot be used to match static features, so we instead apply planar geometric constraints to moving objects tracked throughout the scene. By robustly matching and fitting tracked objects to a planar model, we align the scene's ground plane across multiple views and decompose the planar alignment matrix to recover the 3D relative camera and ground plane positions. We demonstrate this technique in both a controlled lab setting where we test the effects of errors in the intrinsic camera parameters, and in an uncontrolled, outdoor setting. In the latter, we do not assume synchronized cameras and we show that enforcing geometric constraints enables us to align the tracking data in time. In spite of noise in the intrinsic camera parameters and in the image data, the system successfully transforms multiple views of the scene's ground plane to an overhead view and recovers the relative 3D camera and ground plane positions     

基于场景模型的多摄像机目标跟踪算法

在多个相机组成的视频监视系统中,当目标物移出某一相机的视野而进入下一个时,如何实现相机的交接,实现目标物的继续跟踪是监视系统中要解决的关键问题。针对该问题,提出了一种基于位置比较的多摄像机运动目标跟踪方法。为获得目标物的位置,建立多个相机与目标物世界坐标之间映射关系的场景模型,并根据目标物出现在不同相机之间的视野边界线上的瞬间时刻的位置来给出重叠视野的边界线。由此可对任意角度摆放的多个具有重叠视野的相机之间运行的目标物进行接力跟踪。该方法可以适应多个目标物同时进入场景的情况,实验结果表明,该方法具有较高的鲁棒性,能够满足视频跟踪的实时性要求。     

Compression of multiple depth maps for IBR

Image-based rendering techniques include those with geometry and those without. Geometric information in the form of a depth map aligned with the image holds a lot of promise for IBR due to the several methods available to capture it. It can improve the quality of generated views using a limited number of views. Compression of light fields or multiple images has attracted a lot of attention in the past. Compression of multiple depth maps of the same scene has not been explored much in the literature. We propose a method for compressing multiple depth maps in this paper using geometric proxy. Different quality of rendering and compression ratio can be achieved by varying different parameters. Experiments show the effectiveness of the compression technique on several model data.     

基于高斯混合模型和AdaBoost的夜间车辆检测

针对夜间车辆检测精度相对不高的问题，提出通过构建车头灯对空间几何关系的高斯混合模型（GMM）和采用逆投影车辆样本的AdaBoost分类器准确检测夜间车辆的方法。首先，在交通场景中根据车头灯对的空间位置关系设置逆投影面，通过图像预处理粗定位车灯区域；其次，在逆投影图像下利用车头灯对的空间几何关系构建车灯对的高斯混合模型，初步匹配车头灯对；最后，采用逆投影车辆样本，利用AdaBoost分类器进一步准确检测车辆。实验在3个交通场景的检测结果表明，与原始图像下的AdaBoost方法相比，所提方法的检测率提高了1.93%，漏检率降低了17.83%，误检率降低了27.61%；与D-S （Dempster-Shafer）证据理论方法相比，检测率提高了2.03%，漏检率降低了7.58%，误检率降低了47.51%。所提方法提高了相对检测精度，减少了地面反光和影子等的干扰，满足交通场景中夜间车辆检测的可靠性和准确性的要求。     

Automatic Registration of Multi‐Projector Domes Using a Single Uncalibrated Camera

In this paper we present a novel technique for easily calibrating multiple casually aligned projectors on spherical domes using a single uncalibrated camera. Using the prior knowledge of the display surface being a dome, we can estimate the camera intrinsic and extrinsic parameters and the projector to display surface correspondences automatically using a set of images. These images include the image of the dome itself and a projected pattern from each projector. Using these correspondences we can register images from the multiple projectors on the dome. Further, we can register displays which are not entirely visible in a single camera view using multiple pan and tilted views of an uncalibrated camera making our method suitable for displays of different size and resolution. We can register images from any arbitrary viewpoint making it appropriate for a single head‐tracked user in a 3D visualization system. Also, we can use several cartographic mapping techniques to register images in a manner that is appropriate for multi‐user visualization. Domes are known to produce a tremendous sense of immersion and presence in visualization systems. Yet, till date, there exists no easy way to register multiple projectors on a dome to create a high‐resolution realistic visualizations. To the best of our knowledge, this is the first method that can achieve accurate geometric registration of multiple projectors on a dome simply and automatically using a single uncalibrated camera.     

Robust scale invariant target detection using the scale-space theory and optimization for IRST

This paper presents a new small target detection method using scale invariant feature. Detecting small targets whose sizes are varying is very important to automatic target detection in infrared search and track (IRST). The conventional spatial filtering methods with fixed sized kernel show limited target detection performance for incoming targets. The scale invariant target detection can be defined as searching for maxima in the 3D (x, y, and scale) representation of an image with the Laplacian function. The scale invariant feature can detect different sizes of targets robustly. Experimental results with real FLIR images show higher detection rate and lower false alarm rate than conventional methods. Furthermore, the proposed method shows very low false alarms in scan-based IR images than conventional filters.     

A traverse inspection system for high precision visual on-loom fabric defect detection

A self-contained inspection system for vision-based on-loom fabric defect detection is presented in this paper. Design and loom integration of a traversing camera sled, a camera vibration damper and a complementary back-light illumination are presented and discussed. Image acquisition strategies and traverse control are described to complete the discussion on hardware and mechanics. The main part of the paper focuses on a novel algorithmic framework for woven fabric defect detection in highly resolved (1,000+ ppi) image data. Within this scope, single yarns are tracked and measured in terms of position, size, and appearance in real time. An inspection prototype has been mounted onto an industrial loom. Extensive on-line and off-line evaluations for various fabric materials gave precise and stable detection results with few false alarms. A brief cost analysis for the prototype system is provided and completes the presentation of the system.     

基于粒特征和连续Adaboost的人脸检测

提出了一种基于粒特征和连续Adaboost算法的人脸检测方法.它使用粒特征并扩展贝叶斯决策弱分类器,设计具有连续置信度输出的查找表型弱分类器形式,构造出弱分类空间,使用大规模的训练集和验证集,采用连续Adaboost算法学习得到Boosting动态级联型的人脸检测器.在CMU MIT正面人脸测试集上,误报20个时,检测率为90%以上.在一台Pentium Dual 1.2 GHz的PC上,处理一幅大小为320×240像素大小的图片平均需100 ms.实验结果表明该方法取得了比较好的精度和速度.     

Detecting people in dense crowds

We propose a scheme to detect individuals in any image frame of a video sequence showing densely crowded scenes against cluttered backgrounds. The method uses only spatial information, and in an initial pass through the image a trained Viola–Jones-type local detector is used to locate individuals in the densely crowded scene. This yields a large number of false alarms. Hence, in a second step, we seek to reduce the false alarms, and propose two methods for this. In the first, color information from the initially detected windows is passed to a classifier to reduce the false alarms. This classifier consists of a cascade of boosted classifiers with Haar-like features as input and is trained with color information from local windows. In the second method, a weak perspective model of an uncalibrated camera is used to further reduce the false alarm rate while maintaining the detection rate. This is based on the size and locations of the detections in the image frame, without the use of any 3D world information. Results are presented in the form of receiver operating characteristic curves. For instance, at a 79.0% detection accuracy, the false alarm rate is 20.3%.     

Robotic Intelligence for Industrial Automation: Object Flaw Auto Detection and Pattern Recognition by Object Location Searching, Object Alignment, and Geometry Comparison

This research investigates the techniques using image subtraction to find flaws in the cosmetic products. The technique developed in this research moves the perfect image to overlap with the flawed image. Then, the perfect and the flawed images are aligned in the same orientation. After the perfect image has overlapped with the flawed image, the flawed image is subtracted from the perfect image. If there are flaws in the flawed image, after the image subtraction, the flaws will remain in the subtracted result. From beginning to end the inspection is done by machine automatically. There is no further human effort involved. The technique developed in this research can find the flaws in two-dimensional images very accurately. This paper explains the method using the second moment to find the orientations of cosmetic products. By the orientations of the cosmetic products, the perfect cosmetic product and the flawed cosmetic product can be aligned in the same orientation. A detailed process of image rotation is addressed in this paper.     

基于小波和形态学的自然地面背景红外目标检测方法

本文采用小波变换和形态Top-Hat滤波相结合的方法抑制红外图像的混合噪声及大面积的背景干扰,减少图像模糊和增强目标;通过选择适当的结构元素进行系列形态组合运算,去掉虚警点而获得少量候选目标;搜索局部极大值并确定阈值,分离出真正的目标.实验结果表明,该方法能有效检测和分割出低信噪比复杂自然背景红外图像中的目标.     

What you see is what you get [self-calibrating camera lens distortion]

A self-calibration method is presented for self-calibrating camera lens distortion by using only the image correspondences of two views. Two images of a single object are related by the epipolar geometry, which can be described by a 3 /spl times/ 3 singular matrix called fundamental matrix. It captures all geometric information contained in two images. An optimization method is applied to minimize the epipolar distances of the two images by adjusting the camera lens radial distortion coefficient. The merit of the method is that it does not rely on any ground truth data. Simulation and experimental results are given to demonstrate the applicability of the method.     

Intelligent segmentation method for real-time defect inspection system

To extract desired flaws from various types of images, the integration of different segmentation methods is required. In this study, we present an intelligent method for automatic selection of a proper image segmentation method upon detecting a particular flaw type. The new method is capable of choosing the most suitable method from four segmentation methods currently available. The automatic selection procedures start from the pre-segmentation of an image to be examined. Then, the predetermined features are extracted from the original, foreground, and background images. After that, a suitable segmentation method will be selected using a classifier based on six features. Finally, the image is re-segmented by the selected segmentation method to discover flaws. The proposed method has been tested using 1676 defective images. The results show a significant reduction in misclassification rate from about 44% to 13.96%.     

Real-time exploration and photorealistic reconstruction of large natural environments

This paper presents a hybrid (geometry- and image-based) framework suitable for providing photorealistic walkthroughs of large, complex outdoor scenes, based only on a small set of real images from the scene. To this end, a novel data representation of a 3D scene is proposed, which is called morphable 3D panoramas. Motion is assumed to be taking place along a predefined path of the 3D environment and the input to the system is a sparse set of stereoscopic views at certain positions (key positions) along that path (one view per position). An approximate local 3D model is constructed from each view, capable of capturing the photometric and geometric properties of the scene only locally. Then, during the rendering process, a continuous morphing (both photometric as well as geometric) takes place between successive local 3D models, using what we call a ‘morphable 3D model’. For the estimation of the photometric morphing, a robust algorithm capable of extracting a dense field of 2D correspondences between wide-baseline images is used, whereas, for the geometric morphing, a novel method of computing 3D correspondences between local models is proposed. In this way, a physically valid morphing is always produced, which is thus kept transparent from the user. Moreover, a highly optimized rendering path is used during morphing. Thanks to the use of appropriate pixel and vertex shaders, this rendering path can be run fully in 3D graphics hardware and thus allows for high frame rates. Our system can be extended to handle multiple stereoscopic views (and therefore multiple local models) per key position of the path (related by a camera rotation). In this case, one local 3D panorama (per key position) is constructed, comprising all local 3D models therein, and so a ‘morphable 3D panorama’ is now used during the rendering process. For handling the geometric consistency of each 3D panorama, a technique which is based on solving a partial differential equation is adopted. The effectiveness of our framework is demonstrated by using it for the 3D visual reconstruction of the Samaria Gorge in Crete.     

基于Q学习算法的X光主动视觉安检方法

针对主动视觉安检方法检测性能不高和检测速度慢的问题,基于Q学习（QL）算法提出了采用状态回溯的启发式Q学习（HASB-QL）算法进行最佳视角估计。该算法引入代价函数和启发函数,提高了学习效率,加快了Q学习收敛。首先,对通过安检扫描仪获取的X光图像进行单视角检测;然后,对姿势作出估计并通过在状态回溯过程中比较重复动作的选择策略获取最佳旋转角度,再次进行单视角检测,直到检测到危险品;此外,在检测过程中多于一个视角时,建立几何约束以消除误报。对GDXray数据集中的手枪和剃刀刀片的X光图像进行实验,实验结果表明,相比于以Q学习为基础的主动视觉算法,改进的主动视觉算法检测手枪所得精确率和召回率之间的加权平均值F₁值提高了9.60%,检测速度提高了12.45%;检测剃刀刀片所得的F₁值提高了2.51%,速度提高了17.39%。所提算法提高了危险品检测的性能和速度。     

Antifaces: a novel, fast method for image detection

This paper offers a novel detection method, which works well even in the case of a complicated image collection. It can also be applied to detect 3D objects under different views. The detection problem is solved by sequentially applying very simple filters (or detectors), which are designed to yield small results on the multitemplate (hence antifaces), and large results on “random” natural images. This is achieved by making use of a simple probabilistic assumption on the distribution of natural images, which is borne out well in practice. Only images which passed the threshold test imposed by the first detector are examined by the second detector, etc. The detectors are designed to act independently so that their false alarms are uncorrelated; this results in a false alarm rate which decreases exponentially in the number of detectors. The algorithm's performance compares favorably to the well-known eigenface and support vector machine based algorithms, but is substantially faster     

A geometric comparison of algorithms for fusion control instereoscopic HTDs

This paper concerns stereoscopic virtual reality displays in which the head is tracked and the display is stationary, attached to a desk, tabletop or wall. These are called stereoscopic HTDs (head-tracked displays). Stereoscopic displays render two perspective views of a scene, each of which is seen by one eye of the user. Ideally, the user's natural visual system combines the stereo image pair into a single, 3D perceived image. Unfortunately, users often have difficulty fusing the stereo image pair. Researchers use a number of software techniques to reduce fusion problems. This paper geometrically examines and compares a number of these techniques and reaches the following conclusions: In interactive stereoscopic applications, the combination of view placement, scale, and either false eye separation or α-false eye separation can provide fusion control that is geometrically similar to image shifting and image scaling. However, in stereo HTDs, image shifting and image scaling also generate additional geometric artifacts that are not generated by the other methods. We anecdotally link some of these artifacts to exceeding the perceptual limitations of human vision. While formal perceptual studies are still needed, geometric analysis suggests that image shifting and image scaling may be less appropriate than the other methods for interactive, stereo HTDs