Search

Video-based moving vehicle detection and tracking is an important prerequisite for vehicle counting under complex transportation environments. However, in the complex natural scene, the conventional optical flow method cannot accurately detect the boundary of the moving vehicle due to the generation of the shadow. In addition, traditional vehicle tracking algorithms are often occluded by trees, buildings, etc., and particle filters are also susceptible to particle degradation. To solve this problem, this paper proposes a kind of moving vehicle detection and tracking based on the optical flow method and immune particle filter algorithm. The proposed method firstly uses the optical flow method to roughly detect the moving vehicle and then uses the shadow detection algorithm based on the HSV color space to mark the shadow position after threshold segmentation and further combines the region-labeling algorithm to realize the shadow removal and accurately detect the moving vehicle. Improved affinity calculation and mutation function of antibody are proposed to make the particle filter algorithm have certain adaptivity and robustness to scene interference. Experiments are carried out in complex traffic scenes with shadow and occlusion interference. The experimental results show that the proposed algorithm can well solve the interference of shadow and occlusion and realize accurate detection and robust tracking of moving vehicles under complex transportation environments, which has the potentiality to be processed on a cloud computing platform.

The paper deals with the development of the algorithm for processing redundant information in the airborne non-collinear measuring instrument taking into account the possibility of faults. The research is based on such methods as neural networks, and processing information in redundant non-collinear inertial measuring instruments. As a result, the algorithm of redundant information processing assigned for use on the moving vehicle during its operation is developed. The proposed solution is acceptable for unmanned aerial vehicles due to a decrease in the computational burden.

In this work, we present LIBRE: LiDAR Benchmarking and Reference, a first-of-its-kind dataset featuring 10 different LiDAR sensors, covering a range of manufacturers, models, and laser configurations. Data captured independently from each sensor includes three different environments and configurations: static targets, where objects were placed at known distances and measured from a fixed position within a controlled environment; adverse weather, where static obstacles were measured from a moving vehicle, captured in a weather chamber where LiDARs were exposed to different conditions (fog, rain, strong light); and finally, dynamic traffic, where dynamic objects were captured from a vehicle driven on public urban roads, multiple times at different times of the day, and including supporting sensors such as cameras, infrared imaging, and odometry devices. LIBRE will contribute to the research community to (1) provide a means for a fair comparison of currently available LiDARs, and (2) facilitate the improvement of existing self-driving vehicles and robotics-related software, in terms of development and tuning of LiDAR-based perception algorithms., Comment: Accepted for oral presentation at IEEE Intelligent Vehicles Symposium (IV2020), https://2020.ieee-iv.org/ LIBRE dataset available at https://sites.google.com/g.sp.m.is.nagoya-u.ac.jp/libre-dataset/ Reference video available at https://youtu.be/rWyecoCtKcQ

For detecting the number plate from the moving vehicle there are many algorithm has been developed but still this area always remain evolving each every year, Now a days it is very important to apply image processing and machine learning techniques in various field license plate recognition or number plate recognition is widely used application of image processing and machine learning here in this paper we are going to discuss different methodologies used for recognition of number plate and extracting numbers and characters from number plate and also the accuracy reorganization and future scope of different available methods.

Questions of application of laboratory methods of analysis in forensic-medical examination of modern traffic accidents (during the contact of moving vehicle with the pedestrian and injury of the driver and passengers inside the car) are developed in this article. Limitations of laboratory analysis organization, first of all of laboratory medical-criminalistic analysis of bone fragments and soft tissues with injuries, and laboratory analysis of clothes and shoes of the injured persons, what decreases the importance of forensic- medical examination, were represented. To improve this it is offered to apply the algorithm of forensic-medical examination, which includes both medical-criminalistic and forensic-histological analysis

Object detection represents the most important component of Automated Vehicular Surveillance (AVS) systems. Moving vehicle detection based on background subtraction, with fixed morphological parameters, is a popular approach in AVS systems. However, the performance of such an approach deteriorates in the presence of sudden illumination changes in the scene. To address this issue, this paper proposes a method to adjust in real-time the morphological parameters to the illumination changes in the scene. The method is based on machine learning. The features used in the machine learning models are first, second, third and fourth-order statistics of the grayscale images, and the outputs are the appropriate morphological parameters. The resulting background subtraction-based object detection is shown to be robust to illumination changes, and to significantly outperform the conventional approach. Further, artificial neural network (ANN) is shown to provide better performance than Naive Bayes and K-Nearest Neighbours models.

A new methodology for an on-road vehicle motion prediction for several seconds.Our algorithm is operated on a stereo camera system mounted on a moving vehicle.We use a vehicle motion model based on the Gaussian Mixture Model (GMM).The Rapidly-Exploring Random Tree (RRT) allows to consider surrounding environments.We propose a simulation based method to combine the results of the RRT and GMM. The movement of a vehicle is much affected by surrounding environments such as road shapes and other traffic participants. This paper proposes a new vehicle motion prediction method to predict future motion of an on-road vehicle which is observed by a stereo camera system mounted on a moving vehicle. Our proposed algorithm considers not only the history movement of the observed vehicle, but also the environment configuration around the vehicle. To find feasible paths under a dynamic road environment, the Rapidly-Exploring Random Tree (RRT) is used. A simulation based method is then applied to generate future trajectories by combining results from RRT and a motion prediction algorithm modelled as a Gaussian Mixture Model (GMM). Our experiments show that our approach can predict future motion of a vehicle accurately, and outperforms previous works where only motion history is considered for motion prediction.

2017 IEEE Sensors In this paper we present the design and the first results of a prototype road condition sensor based on diffuse reflectance spectroscopy in the near infrared using semiconductor lasers. The objective of the sensor is to alert drivers or autonomous driving assistance systems of a road condition, such the presence of water or mud or ice, that will reduce the surface grip a few meters ahead the vehicle. This work was supported by the Spanish Ministry of Economy and Competitiveness under the RTC-2015-4205-4 “SALICAR” grant.

We present a vision system that automatically predicts yield in vineyards accurately and with high resolution. Yield estimation traditionally requires tedious hand measurement, which is destructive, sparse in sampling, and inaccurate. Our method is efficient, high-resolution, and it is the first such system evaluated in realistic experimentation over several years and hundreds of vines spread over several acres of different vineyards. Other existing research is limited to small test sets of 10 vines or less, or just isolated grape clusters, with tightly controlled image acquisition and with artificially induced yield distributions. The system incorporates cameras and illumination mounted on a vehicle driving through the vineyard. We process images by exploiting the three prominent visual cues of texture, color, and shape into a strong classifier that detects berries even when they are of similar color to the vine leaves. We introduce methods to maximize the spatial and the overall accuracy of the yield estimates by optimizing the relationship between image measurements and yield. Our experimentation is conducted over four growing seasons in several wine and table-grape vineyards. These are the first such results from experimentation that is sufficiently sized for fair evaluation against true yield variation and real-world imaging conditions from a moving vehicle. Analysis of the results demonstrates yield estimates that capture up to 75% of spatial yield variance and with an average error between 3% and 11% of total yield.

In this paper, we propose the spatial domain based vehicle detection scheme. This proposed scheme combines the Sobel edge detection method with background subtraction in YCbCr color space. The scheme detects the vehicle in Y(Luminance), Cb and Cr (chrominance) components of the vehicle image using background subtraction and combines the three images. Edge detection method determines edge information of the luminance component of the vehicle image. The image combined with edge detection and background difference is implemented filling and filtering operation. The robustness of the proposed scheme is analyzed considering different types of vehicle image.

Moving vehicle detection based on video processing has been widely used in intelligent transportation system recently. However there are also many problems, such as dynamic background, ghost region, and shadow of moving objects. This paper proposes an improved ViBe object detection algorithm. First, an accurate background image is obtained by using the multi-frame averaging method, and then the background model is initialized by this accurate background image, thus effectively reducing the generation of ghost region. Whenever there is no moving object for a fixed number of consecutive frames in the video, this frame is updated to the background image. Conservative update strategy and foreground point counting method are adopted to update the background and reduce the impact of dynamic background on the foreground detection. Next, the foreground image detected by improved ViBe algorithm is input into the shadow elimination method proposed in this paper. Shadows in foreground pixels are detected in RGB color space, and then the pixels determined as shadows are eliminated. Finally, accurate moving vehicles are obtained. Our algorithm can effectively eliminate the shadows of moving vehicles, quickly adapt to background changes and illumination mutation, and get accurate moving objects, which is helpful for vehicle contour extraction and subsequent image processing.

License plate recognition systems can be classified into several categories: systems with single camera for motionless vehicle, systems with single camera for moving vehicle, and systems with multiple cameras for moving vehicles on highways (one camera for each lane). In this paper we present an innovative system which can locate multiple moving vehicles and recognize their license plates with only one single camera. Obviously, our system is highly cost effective in comparison with other systems. Our system has license plate localization success rate 94% and license plate recognition success rate 88%. These success rates are pretty satisfiable considering the system is working on fast moving vehicles on highway.

Autonomous Vehicles require accurate environment information in order to perform scene understanding with very high accuracy. For this reason, we must design perception algorithms that are robust and must perform in any environmental and weather conditions. Since fog is one of the most dangerous weather phenomena for driving we propose an embedded solution that is able to restore the contrast of images captured from a moving vehicle in fog conditions. Our solution is based on a state of the art contrast restoration algorithm and is able to provide high quality images in real time. The visibility in the enhanced images is greatly improved such that other image processing algorithms can be applied on the resulting images. We have chosen the Xilinx ZYNQ embedded platform since it provides the best tradeoff in terms of cost, performance and form-factor.

The paper presented a vehicle speed estimation algorithm based on moving target detection in video surveillance. Firstly, the features of moving vehicles were extracted by using the three frame difference method and the background difference method; secondly, tracked and positioned the moving target according to moving vehicle centroid feature extraction method; finally, the vehicle speed was estimated based on mapping relationship between the pixel distance with the actual distance.

This paper describes the tracking guidance for the autonomous UAV landing and the vision-based detection of the marker on the moving vehicle with the real-time image processing system. The detection of the marker is based on the color detection algorithm with morphological filter and the tracking guidance is based on the relative distance between the UAV and the vehicle. In order to track and land on the moving vehicle, we have built the integrated system on the UAV platform that conducts the landing mission with the on-board image processing system. The flight test was performed with the moving vehicle in the outdoor environment, and we validated the algorithm for the vision-based autonomous landing.

Moving vehicle detection from unmanned aerial vehicles (UAV) is becoming an increasingly important research topic in traffic monitoring, surveillance and military applications. Owing to the low-quality of UAV videos and the movement of the platform, vehicle detection is a challenging task. Existing algorithms that are generally designed for stationary cameras are ineffectively under this situation. This paper proposes an accurate moving vehicle detector. Significant contributions include a real-time, high detection rate approach for motion detection with image registration, candidate targets detection and vehicle screening using convolutional neural network. Experiments on a variety of data sets show the successful detection of moving vehicle under varying conditions. Currently the detection rate for vehicle is up to 90% and the average false alarm rate is less than 10%.

We address the problem of moving vehicle detection for dynamic traffic environment with moving cameras. The proposed system obtains visual data from cameras mounted on a moving host vehicle. The optical flow method is employed to estimate the movement of the camera and the relative movement of the background. Moving vehicles are detected by difference of velocity to the background movement model.

As one of the important topics in computer vision, moving vehicle segmentation has attracted considerable attention of researchers. However, robust detection is hampered by the interferential moving objects in dynamic scenes. In this paper, we address the problem of the moving vehicles segmentation in the dynamic scenes. Based on the distinct motion property of the dynamic background and that of the moving vehicles, we present an adaptive motion histogram for moving vehicles segmentation. The presented algorithm consists of two procedures: adaptive background update and motion histogram-based vehicles segmentation. In the adaptive background update procedure, we make use of the lighting change of the scene and present a novel method for background evolving. In the motion histogram-based vehicles segmentation procedure, an adaptive motion histogram is maintained and updated according to the motion information in the scenes, and the moving vehicles are then detected according to the motion histogram maintained. Experimental results of typical scenes demonstrate robustness of the proposed method. Quantitative evaluation and comparison with the existing methods show that the proposed method provides much improved results.

This paper presents a license plate recognition system of moving vehicles for Egyptian license plates. The proposed system extracts the characters features of a plate from a captured image by a digital video camera. Then it identifies numbers and characters of the plate using a multilayer neural network. The proposed system consists mainly of two processes: the first one is the training process, in which a database is build and the neural networks used for recognition are trained. The second one is the recognition process which consists of vehicle detection from the stream of frames, a plate position estimation, words and numbers segmentation, and plate recognition stages.

In this paper, we propose a fast and accurate scene recognition system from moving vehicle, which is a combination of Higher-order Local Auto-Correlation (HLAC) features and the linear discriminant analysis. The scene recognition system of moving vehicles should classify successive images into categories under limited computational resources, have robustness to cope with environmental changes, and adopt various scenes, which do not have explicit features. Our algorithm satisfies those requirements. We evaluated our algorithm in intersection and highway recognition experiments, and showed that our algorithm not only has comparable recognition accuracy compared with the state-of-the-art methods but also outperforms those methods in computational speed.

Detecting vehicles in video is a challenging problem owing to the motion of vehicles, the camera and the background and to variations of speed. This paper proposes a classifier based supervised method to detect moving vehicles from a moving camera. Dense scale invariant feature transform (dense SIFT) descriptors are used as features to describe the pattern of the object. And Extreme Learning Machine provides excellent generalization performance at fast speed. Our sample images taken by a camera in helicopter include 2000 images. Experiment results shown that this proposed method has not only good overall performance but also low computational cost.

Vehicle speed plays a crucial role in determining safety in traffic. In this paper vehicle speed is automatically estimated from video sequences taken from a single calibrated camera, so that this information can be used in controlling the traffic. The vehicle motion is detected and tracked along the frames. Our proposed system tracks the vehicles and gives the estimated speed of the vehicles using optical flow technique-Lucas Kanade using Pyramidal implementation.

There are some defects in traditional three-frame differencing algorithm for moving vehicle detection, such as intermittent moving region, low detection rate of dark moving objects, multi-target cross and target splitting. This paper proposes an object reconstruction algorithm for moving vehicle detection based on three-frame differencing (ORTFD). The linear point operation on the original video image is employed to reduce the dispersion of dark moving vehicles. Three-frame differencing operation is used to get the contours of moving vehicles, and perform morphological processing on the binary image is adopted to exclude non target objects, also to improve the integrity of the obtained contours. Finally, progressive scan on the whole binary image is taken to find out connected region surrounded by the contour, a preset-threshold-based region filling method (PTRF) is designed to reconstruct the moving object. Algorithm simulation results show that ORTFD improve the detection rate of dark moving vehicle object, and accelerate the detection speed, the precision ratio of moving vehicle detection has been also improved. ORTFD algorithm can detect and recognize vehicles in real-time with high-accuracy for multiple targets under complex traffic conditions.

Detecting vehicle in video is a challenging problem owing to the motion of vehicle, camera and background, and to variations of speed. This paper proposes a classifier-based supervised method to detect moving vehicle from a moving camera. Motion descriptors based on oriented histograms of differential optical flow be able to describe the motion pattern of the object from changing background. And Extreme Learning Machine provides excellent generalization performance at fast speed. Our sample images taken by a camera in helicopter include 2000 images. Experiment results shown that this proposed method has not only overall discrimination performance but also low computational cost.

In this paper we propose a simple and lightweight solution to estimate the geospatial trajectory of a moving vehicle from images captured by a cellphone exploiting the map and the imagery provided by Google Streetview.

In this paper, we will present a technique for measuring visibility distances under foggy weather conditions using a camera mounted onboard a moving vehicle. Our research has focused in particular on the problem of detecting daytime fog and estimating visibility distances; thanks to these efforts, an original method has been developed, tested and patented. The approach consists of dynamically implementing Koschmieder's law. Our method enables computing the meteorological visibility distance, a measure defined by the International Commission on Illumination (CIE) as the distance beyond which a black object of an appropriate dimension is perceived with a contrast of less than 5%. Our proposed solution is an original one, featuring the advantage of utilizing a single camera and necessitating the presence of just the road and sky in the scene. As opposed to other methods that require the explicit extraction of the road, this method offers fewer constraints by virtue of being applicable with no more than the extraction of a homogeneous surface containing a portion of the road and sky within the image. This image preprocessing also serves to identify the level of compatibility of the processed image with the set of Koschmieder's model hypotheses.

In this article, a method is presented for detecting vehicles in image sequences without prior knowledge about the position of the road. A single camera placed in a moving vehicle provides image data. For the detection of midrange and distant vehicles, a combination of three clues is used: shadow, entropy, and horizontal symmetry. To detect passing vehicles, a temporal differencing and projected motion is used. The algorithms are tested by means of many different experiments. These experiments illustrate the robust and accurate performance of both approaches.

Moving vehicle detection remain very critical and thus intended for Video-based solution, comparing to other techniques and by considering the traffic video sequence recorded from a video camera, this paper presents a video-based solution applied with adaptive subtracted background technology in combination with virtual detector and blob tracking technologies. This paper provides Experimental results moving vehicle detection which is implemented in Visual C++ code with OpenCV, thus the proposed method used for detection.

Moving objects play a key role for gaining scene understanding in aerial surveillance tasks. The detection of moving vehicles can be challenging due to high object distance, simultaneous object and camera motion, shadows, or weak contrast. In scenarios where vehicles are driving on busy urban streets, this is even more challenging due to possible merged detections. In this paper, a video processing chain is proposed for moving vehicle detection and segmentation. The fundament for detecting motion which is independent of the camera motion is tracking of local image features such as Harris corners. Independently moving features are clustered. Since motion clusters are prone to merge similarly moving objects, we evaluate various object segmentation approaches based on contour extraction, blob extraction, or machine learning to handle such effects. We propose to use a local sliding window approach with Integral Channel Features (ICF) and AdaBoost classifier.

Vehicle speed detection is used to estimate the velocity of the moving vehicle using image and video processing techniques. Without any camera calibrations video is captured and analyzed for speed in real time. By employing frame subtraction and masking techniques, moving vehicles are segmented out. Speed is calculated using the time taken between frames and corner detected object traversed in that frames. Finally frame masking is used to differentiate between one or more vehicles. With an average error of +/-2 km/h speed detection was achieved for different video sequences.

In this paper, we propose a method to detect moving obstacles from image data be given from two cameras on the moving vehicle. The moving vehicle is equipped with two cameras for quick detection of moving objects. One (is called fixed camera) is fixed on the front part of the vehicle, and the other (is called movable camera) is able to move within the deck of the vehicle. In this system, the position of the movable camera traces that of the fixed camera. Using this system, it is able to detect moving objects quickly as the vehicle moving.

The automatic detection of visually salient information from abundant video imagery is crucial, as it plays an important role in surveillance and reconnaissance tasks for Unmanned Aerial Vehicle (UAV). A real-time approach for the detection of salient objects on road, e.g. stationary and moving vehicle or people, is proposed, which is based on region segmentation and saliency detection within related domains. Generally, the traditional method specifically depends upon additional scene information and auxiliary thermal or IR sensing for secondary confirmation. However, this proposed approach can detect the interesting objects directly from video imagery captured by optical camera fixed on the small level UAV platform. To validate this proposed salient object detection approach, the 25 Hz video data from our low speed small UAV are tested. The results have demonstrated the proposed approach performs excellently in isolated rural environments.

License plate location (LPL) from moving vehicle images is a key step in license plate recognition (LPR) system. A novel LPL algorithm is proposed in this paper to jointly consider color and texture distribution features of license plate area (LPA) in an image. The method is divided into two steps of detecting vehicles through background subtraction method and locating LPA by combining texture features and color features respectively. Experimental results show that this method can efficiently overcome the limitation of vehicles with complicated background, and can consequently achieve high detection rate in complex situations.

In recent years intelligent transportation systems has been an active research area in computer vision. This work aims to classify vehicles as car, van or truck by using videos which are taken from an uncalibrated camera which sees any highway from above. In this work, moving blobs are detected by a background-foreground extraction algorithm, classification is done by extracting the blob features and histogram of oriented gradients features of detected blobs and by giving those features to support vector machine classifier. Suggested method has been tried on 2 different traffic video and results have been presented.

We present a novel approach for pedestrian detecting on moving vehicle which equipped with low-cost cameras. Our approach is working in a framework which combines two-dimensional human body characteristics and three-dimensional information such as parallax and distance. By constructing a SPM (surface parallax map), it calculates parallax of object which do not belong to the road plane such as human body and obstacles. After recording the scores of all road area, an occlusion image is created, in which high density area indicates people's most likely appearance. Then a SVM (support vector machine) classifier is trained to classify pedestrian and non-pedestrian windows in candidate area. We also propose an algorithm to maintain SPM in real time. We evaluate our approach on real data which are taken from crowded city areas, the efficient and accurate results are demonstrated.

This paper introduces the Comprehensive utilization of symmetric difference methods and background subtraction[1, 2] for moving vehicle detection. Firstly build a reliable background update model, and then combine the basically accurate foreground image got from background subtraction with difference image got from symmetric difference. Finally get the reliable image of moving target, so that the mobile vehicles can be detected from video sequence.

Due to problems of blank holes created by traditional frames difference in moving object detection, this paper develop a method of moving vehicle detection, which is based on three-frame-difference and union set. First, we difference the three grayed images and enhance them. Second, dichotomize images and filter some interference. Third, make the union set operation. Finally, detect the vehicle with morphology process and connectivity analysis. The results of experiment show that this method is characterized by small calculating amount, fast speed, accurate detection and a good robustness.

We present a multi-view stereo method that avoids producing hallucinated surfaces which do not correspond to real surfaces. Our approach to 3D reconstruction is based on the minimal s-t cut of the graph derived from the Delaunay tetrahedralization of a dense 3D point cloud, which produces water-tight meshes. This is often a desirable property but it hallucinates surfaces in complicated scenes with multiple objects and free open space. For example, a sequence of images obtained from a moving vehicle often produces meshes where the sky is hallucinated because there are no images looking from the above to the ground plane. We present a method for detecting and removing such surfaces. The method is based on removing perturbation sensitive parts of the reconstruction using multiple reconstructions of perturbed input data. We demonstrate our method on several standard datasets often used to benchmark multi-view stereo and show that it outperforms the state-of-the-art techniques .

Detecting and localizing other traffic participants, especially pedestrians, from a moving vehicle have many applications in smart vehicles. In this work, we address these tasks by utilizing image sensors, namely stereo cameras mounted on a vehicle. Our proposed method integrates appearance-based pedestrian detection and sparse depth estimation. To benefit from depth estimation, we map the prior distribution of a human’s actual height onto the image to update the detection result. Simultaneously, the depth information that contributed to correct pedestrians’ hypotheses is used for a better localization. The difference with other previous works is that we take the trade-off between accuracy and computational cost in the first place of consideration and try to make the most efficient integration for onboard applications.

The detection of moving object in dynamic case especially outdoor case is very challenging for computer vision. One of the problems in dynamic case is detection of shadow in moving object. This paper presents the improvement of moving vehicle detection using removal shadow technique in RGB color space and frame difference technique. In RGB color space, the luminance and chrominance value are being tuned to remove shadow. The result shows about 75% of shadow can be removed from image moving object detection. The result also shows the improvement of moving object detection.

This work describes a visual odometry by minimizing a quadrifocal constraint in order to calculate the position of an autonomous vehicle (Cycab). This method can also reject outliers corresponding to moving objects such as occlusions and illumination changes. An iterative non-linear trajectory estimation approach is formulated and it is based on a quadrifocal relationship between the image intensities within adjacent views of the stereo pair. The technique is applied to recovering the trajectory of a moving vehicle in a video sequence.

In this paper, we present a novel approach for detection of moving vehicles in traffic videos. We propose a feature-based (corner-based) tracking to track and classify moving vehicles from the extracted ghost or cast shadow. The corner points of the vehicles are detected, labeled and grouped to generate a unique label per vehicle. This approach is able to deal with different types of deformations on the shape of the vehicles due to changes in size, direction and viewpoint. Also, the proposed method is totally free from motion estimation. To demonstrate the robustness and accuracy of our system, the results of the experiments are conducted on traffic videos including different complex background, illumination, motion, camera position, clutter and direction of the vehicles taken from outdoor boulevards and city roads. We detect moving vehicles on an average of 98.8% in a scene. The results show the robustness of our proposed algorithm.

In this paper, Adopt a way that combines with Mean Shift algorithm and Kalman filter to tracking moving vehicle in the paper. At first, Using inter-frame difference algorithm to extract aimed-vehicle. After the aimed-vehicle is processed by binarization and mathematics morphology, we adopt Kalman filter to predict the position of vehicle. Then we adopt Mean shift algorithm to iterate and compute the best position and track. We take the current best position as Kalman filter's observed value to predict next frame image. The experiment results show that, the algorithm can tracking moving vehicle precisely and real-time, and also has better robustness.

In machine vision and the vehicle recognition system? removal of moving vehicle shadows is a significant topic. In this paper, we propose a novel method to detect shadows in traffic video sequences. Firstly, a set of moving regions are segmented from the video sequence using a background subtraction technique. Secondly, the fast normalized cross-correlation (FNCC) is adopted to detect shadows in moving regions from grayscale video sequences. By utilizing three sum-table schemes, the FNCC algorithm dramatically reduces the computational complexity compared to the traditional normalized cross correlation (NCC) algorithm. And our experimental results demonstrate that the proposed shadows removal method is accurate and efficient.

We propose a method of real-time moving vehicle detection using infrared thermal images. It can detect moving vehicles robustly even for bad environments compared with conventional vehicle detection methods using visible light cameras. It can also measure the size of each vehicle around the clock. The algorithm we propose for this detection is designed for a high-speed processing without complicated calculations and also designed for a real-time vehicle detection by using a general-purpose personal computer. Experimental results of our method by use of infrared thermal traffic images both at daytime with vehicular shadows and at nighttime in darkness show that the vehicle detection accuracy is 100%.

A stereo algorithm for detecting and locating obstacles raised above a ground plane is described. It is based on applying a simple multi-model approach using hypothesis tests to the grey levels of preprocessed stereo images. The algorithm has been successfully applied to real images captured from a moving vehicle to recover putative obstacle locations.

Precise relative navigation between an unmanned vehicle and a target, which could be a stationary or a second moving vehicle, is an important capability that has many possible applications including formation operation, as well as autonomous rendezvous and docking of either spacecraft or aircraft. A test bed setup is described where unmanned ground vehicles are used to simulate the physical motion of aerospace vehicles, and provide the attached sensor packages with realistic relative motion in both indoor and outdoor environments. Using an unmanned robotic ground vehicle, the visual servoing problem is investigated and simulated using actual hardware in real world test conditions. A simple camera is used to measure the relative position, orientation, and motion between the vehicle and the target. Color statistical pressure snake algorithms are employed to track a visual target within the camera images in real time. The effectiveness of the nonlinear visual servoing algorithm is demonstrated through experimental hardware tests.

In this paper we present a real time pedestrian detection system that works on low quality infrared videos. We introduce probabilistic templates to capture the variations in human shape, especially for the case where contrast is low and body parts are missing. We present experimental results on infrared videos taken from a moving vehicle in various urban street scenarios to demonstrate the feasibility of the approach.

In this paper, color invariant co-occurrence features for moving vehicle tracking in a known environment is proposed. It extracts moving areas shaped on objects in Web video sequences captured by the Web camera and detects tracks of moving objects. Color invariant co-occurrence matrices are exploited to extract the plausible object blocks and the correspondences between adjacent video frames. The measures of class separability derived from the features of co-occurrence matrices are used to improve the performance of tracking. The experimental results are presented.

We present a new method for obstacle and pedestrian detection with lower computation complexity and higher detection performance on a small target. By calculating stereo disparity according to the road parameters, our system efficiently detects objects above the ground. Experiments on over 10,000 images captured in urban areas demonstrate our method's effectiveness. We also propose a multi-scale compatible pedestrian detector. Detection speed is improved by avoiding size adjustment of input. Experiments on three public pedestrian databases and our HENU database show that our detector achieves better results than the state-of-the-art detection quality at equal speed, especially on pedestrian target with small size.