Video-based Surveillance Research Articles

A maneuver indicator and ensemble learning-based risky driver recognition approach for highway merging areas

Abstract Due to the complex traffic characteristics in highway merging areas, drivers tend to exhibit high-risk driving behaviors. To address the characteristics of driving behavior in highway merging areas, we have developed a real-time identification model for risky drivers by combining a driver risk level labeling method with load balancing-ensemble learning (LB-EL). In this paper, we explore four types of maneuver indicator indexes (MIIs)—acute direction, stomp pedal, dangerous following, and dangerous lane changing—that can describe the negative behaviors of both individual vehicles and vehicle platoons in highway merging areas. To quantize the label driver risk level, we use the interquartile range (IQR) method and Criteria Importance Though Intercriteria Correlation (CRITIC), while we evaluate the reliability of the MII using spatial analysis. Furthermore, we balance the dataset using three load balancing (LB) algorithms and create nine ensemble strategies by pairing adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), and light gradient boosting machine (LGBM) with the three LB algorithms. Finally, we validate the proposed model using trajectory data extracted from UAV videos. The results indicate that the distribution laws of risky driving behaviors in the acute direction and stomp pedal show a high degree of similarity and good matching with the distribution laws of traffic conflict points in existing research. Moreover, the SMOTE-LGBM ensemble model achieves the best performance, reaching an accuracy rate of 93.4% and a recall rate of 92.1%, which demonstrates the validity of our proposed model. This model can be widely applied to recognize risky drivers in video-based surveillance systems.

Black Spot Syndrome in ocean surgeonfish: using video-based surveillance to quantify disease severity and test environmental drivers

Black Spot Syndrome in ocean surgeonfish: using video-based surveillance to quantify disease severity and test environmental drivers

Intelligent Technique for Moving Object Detection from Problematic Video Captured through Camera Sensor

Aim: The significant aim of the proposed work is to develop an adaptive method to compute the threshold during run-time and update it adaptively for each pixel in the testing phase. It classifies motion-oriented pixels from the scene for moving objects using background subtraction and enhances using post-processing. Background: According to the huge demand for surveillance system, society is looking towards an intelligent video surveillance system that detect and track moving objects from video captured through a surveillance camera. So, it is very crucial and highly recommended throughout the globe in numerous domains such as video-based surveillance, healthcare, transportation, and many more. Practically, this research area faces lots of challenging issues such as illumination variation, cluttered background, camouflage, etc. So, this paper has developed an adaptive background subtraction method to handle such challenging problems. Objective: To focus and study the problematic video data captured through the camera sensor. To handle challenging issues available in real-time video scenes. To develop a background subtraction method and update the background model adaptively for moving object detection. Method: The proposed method has been accomplished using the following sections: Background model construction Automatic generation of threshold Background subtraction Maintenance of background model Results: The qualitative analysis of the proposed work is experimented with publicly available datasets and compared with considered state-of-the-art methods. In this work, library sequence (thermal data) of CDNET and other color video frame sequences Foreground aperture, Waving Tree and Camouflage are considered from Microsoft’s Wallflower. The quantitative values depicted in Table- 1. This work demonstrate the better performance of the proposed method as compared to state-ofthe- art methods. It also generates better outcomes and handles the problem of a dynamic environment and illumination variation. Conclusion: Currently, the world is demanding computer vision-based security and surveillancebased applications for society. This work has provided a method for the detection of moving information using an adaptive method of background subtraction approach for moving object detection in video scenes. The performance evaluation depicts better average results as compared to considered peer methods.

Memory-Efficient Discrete Cosine Transform Domain Weight Modulation Transformer for Arbitrary-Scale Super-Resolution

Recently, several arbitrary-scale models have been proposed for single-image super-resolution. Furthermore, the importance of arbitrary-scale single image super-resolution is emphasized for applications such as satellite image processing, HR display, and video-based surveillance. However, the baseline integer-scale model must be retrained to fit the existing network, and the learning speed is slow. This paper proposes a network to solve these problems, processing super-resolution by restoring the high-frequency information lost in the remaining arbitrary-scale while maintaining the baseline integer scale. The proposed network extends an integer-scaled image to an arbitrary-scale target in the discrete cosine transform spectral domain. We also modulate the high-frequency restoration weights of the depthwise multi-head attention to use memory efficiently. Finally, we demonstrate the performance through experiments with existing state-of-the-art models and their flexibility through integration with existing integer-scale models in terms of peak signal-to-noise ratio (PSNR) and similarity index measure (SSIM) scores. This means that the proposed network restores high-resolution (HR) images appropriately by improving the image sharpness of low-resolution (LR) images.

Online Video Anomaly Detection.

With the popularity of video surveillance technology, people are paying more and more attention to how to detect abnormal states or events in videos in time. Therefore, real-time, automatic and accurate detection of abnormal events has become the main goal of video-based surveillance systems. To achieve this goal, many researchers have conducted in-depth research on online video anomaly detection. This paper presents the background of the research in this field and briefly explains the research methods of offline video anomaly detection. Then, we sort out and classify the research methods of online video anomaly detection and expound on the basic ideas and characteristics of each method. In addition, we summarize the datasets commonly used in online video anomaly detection and compare and analyze the performance of the current mainstream algorithms according to the evaluation criteria of each dataset. Finally, we summarize the future trends in the field of online video anomaly detection.

Understanding the risks of China-made CCTV surveillance cameras in Australia

ABSTRACT In the global interconnected economy, China-made information-collecting technologies such as closed-circuit television (CCTV) surveillance cameras have become popular products for routine video-based surveillance. Hikvision and Dahua are the two largest global suppliers of CCTV cameras, with both companies supplying their products to over 200 countries. Despite their popularity, national security concerns are commonly cited when adopting these cameras, citing manufacturer links with the Communist Party of China (CPC), cybersecurity vulnerabilities, and sales recorded in the Xinjiang region, that has records of human rights violations. This paper is structured in three parts: first, we explore the predominance of China-made information-gathering technologies in Australia; second, we summarise common national security concerns usually associated with China-based technology manufacturers; and third, we propose regulatory measures to regulating China-made CCTV cameras in Australia. The paper suggests that while state and Federal decision-makers are free to remove Chinese CCTV surveillance cameras, they should avoid overt politisation. Overall, a stronger focus should be placed on evaluating cybersecurity risks of Internet of Things (IoT) information-collecting technologies and considering their timely and effective regulation from the perspective of individual and national interests.

Gun Detection in CCTV Images using HAAR-Like Features

Automated video-based surveillance is an important area of research to assist the security personnel to detect the incident of any abnormal events in the surroundings. The objective of this paper is to develop a framework for automatic gun detection using closed-circuit television (CCTV) images. The methodology presented in this paper involves the development of a framework for automatic gun detection using closed-circuit television (CCTV) images, with the aim of enhancing the surveillance of crime and improving human security. The proposed approach consists of a dataset of CCTV images containing instances of guns, as well as non-gun images for comparison. These images would be used to train the proposed algorithm to recognize and identify guns in future CCTV images. The proposed framework is designed for an indoor environment and uses Haar-like features for gun detection. The proposed system involves the installation of CCTV cameras in a suitable corner of an indoor environment for surveillance. The CCTV cameras capture the scene and the frames of the scene are compared with a predefined dataset for automatic gun detection. The proposed approach draws a bounding box and raises an alarm if it detects a gun in a frame extracted from a captured scene. This provides a visual indication of the presence of a gun, making it easier for relevant authorities to quickly identify and respond to the threat. The proposed system shows promising results in real-time applications and about 90% accuracy has been achieved.

Automated Video-Based Air Traffic Surveillance System for Counting General Aviation Aircraft Operations at Non-Towered Airports

The vast majority of U.S. airports are not equipped with control towers, which limits their ability to keep records of flight operations. Attempts have been made to use sensor-based technologies to count aircraft operations at non-towered airports; however, they exhibit limited accuracy. To this end, we developed an automated video-based surveillance system capable of detecting general aviation aircraft departure and landing operations, which comprise the vast majority of operations at non-towered airports. The proposed computer vision method is comprised of three modules: aircraft detection, aircraft tracking, and operations count and classification. We explored different camera layouts and state-of-the-art machine learning and deep learning methods to determine the best settings to extract operations trajectory features for operations count and classification. The proposed method was tested at five non-towered airports. Integrating deep-neural-network-based aircraft detectors and image-correlation-based aircraft trackers achieved an accuracy of about 95%, while ensuring processing times that are needed for real-time implementation.

Using video-based surveillance for monitoring hand hygiene compliance according to the World Health Organization (WHO) Five Moments framework: A pragmatic trial.

To examine the utility of video-based monitoring systems (VMSs) for auditing hand hygiene compliance according to the World Health Organization (WHO) Five Moments. Pragmatic quasi-experimental observation trial. The New South Wales Biocontainment Centre, Westmead, New South Wales, Australia. Volunteer healthcare workers (HCWs). Six high-fidelity simulations were recorded and subsequently assessed for their ability to audit hand hygiene compliance according to the WHO Five Moments for hand hygiene criteria using tools provided by the National Hand Hygiene Initiative (NHHI). In total, 206 minutes of recorded footage were reviewed in 120 minutes, yielding 111 moments. Overall HCW hand hygiene compliance was 88% according to the WHO Five Moments framework. The cost per moment was $0.91 AUD ($0.66 USD) and the time required per moment was 64 seconds. Auditing of hand hygiene compliance according to all 5 of the WHO Five Moments from recorded footage is not only possible but provides cost and time savings. In addition, the process may produce output that is less subject to the biases inherent in direct human observational auditing.

Risky Driving Behavior Recognition Based on Vehicle Trajectory.

This paper proposes a measurement of risk (MOR) method to recognize risky driving behavior based on the trajectory data extracted from surveillance videos. Three types of risky driving behavior are studied in this paper, i.e., speed-unstable driving, serpentine driving, and risky car-following driving. The risky driving behavior recognition model contains an MOR-based risk evaluation model and an MOR threshold selection method. An MOR-based risk evaluation model is established for three types of risky driving behavior based on driving features to quantify collision risk. Then, we propose two methods, i.e., the distribution-based method and the boxplot-based method, to determine the threshold value of the MOR to recognize risky driving behavior. Finally, the trajectory data extracted from UAV videos are used to validate the proposed model. The impact of vehicle types is also taken into consideration in the model. The results show that there are significant differences between threshold values for cars and heavy trucks when performing speed-unstable driving and risky car-following driving. In addition, the difference between the proportion of recognized risky driving behavior in the testing dataset compared with that in the training dataset is limited to less than 3.5%. The recognition accuracy of risky driving behavior with the boxplot- and distribution-based methods are, respectively, 91% and 86%, indicating the validation of the proposed model. The proposed model can be widely applied to risky driving behavior recognition in video-based surveillance systems.

Gait recognition using a few gait frames.

Gait has been deemed as an alternative biometric in video-based surveillance applications, since it can be used to recognize individuals from a far distance without their interaction and cooperation. Recently, many gait recognition methods have been proposed, aiming at reducing the influence caused by exterior factors. However, most of these methods are developed based on sufficient input gait frames, and their recognition performance will sharply decrease if the frame number drops. In the real-world scenario, it is impossible to always obtain a sufficient number of gait frames for each subject due to many reasons, e.g., occlusion and illumination. Therefore, it is necessary to improve the gait recognition performance when the available gait frames are limited. This paper starts with three different strategies, aiming at producing more input frames and eliminating the generalization error cause by insufficient input data. Meanwhile, a two-branch network is also proposed in this paper to formulate robust gait representations from the original and new generated input gait frames. According to our experiments, under the limited gait frames being used, it was verified that the proposed method can achieve a reliable performance for gait recognition.

An Efficient Approach for Detecting and Classifying Moving Vehicles in a Video Based Monitoring System

Moving objects detection, type recognition, and traffic analysis in video-based surveillance systems is an active area of research which has many applications in road traffic monitoring. This paper is on using classical approaches of image processing to develop an efficient algorithm for computer vision based on traffic surveillance system that can detect and classify moving vehicles, besides serving some other traffic analysis issues like finding vehicles speed and heading, tracking specified vehicles, and finding traffic load. The algorithm is designed to be flexible for modification to fulfill the changes in design objectives, having limited computation time, giving good accuracy, and serves inexpensive implementation. A 92% of success is achieved for the considered test, with the missed cases being abnormal that are not defined to the algorithm. The computation time, with a platform (hardware and software) dependent, the algorithm took to produce results was parts of milliseconds. A CNN based deep learning classifier was built and evaluated to judge the feasibility of involving a modern approach in the design for the targeted aims in this work. The modern NN based deep learning approach is very powerful and represents the choice for many very sophisticated applications, but when the purpose is restricted to limited requirements, as it is believed the case is here, the reason will be to use the classical image processing procedures. In making choice, it is important to consider, among many things, accuracy, computation time, and simplicity of design, development, and implementation.

Proposing SVM and HOG Techniques for Effective Face Recognition in Video Surveillance

Face Recognition is an active topic among Machine Learning Researchers for two decades owing to its increasing demand in security monitoring applications. The present Techniques while being working has some constraints. The challenges emerge with the orientation, quality, and expression, variations in lightning, or facial occlusions, which has a direct impact on the facial captures using video-based surveillance. This results in performance and accuracy issues. The current surveillance applications require more computational complexity with less accuracy and performance. The proposed video surveillance system overcomes these limitations of existing systems and provides maximum effective security with minimum computational complexity. The proposed Video security monitoring system provides a complete face localization, detection, and recognition. The draw out facial image data is compared with facial dataset images. The facial data is obtained from the video dataset accessed from the real environment. The face image is authenticated if a match is found and is declared unauthenticated otherwise. The security alarm after the unauthenticated alerts the security personal for further action. Hence, the proposed system is more non-evasive, accurate and reliable.

Electric Gadgets Surveillance and Controlling Systems Based on Internet of Things IoT

Over the last few a long time, brilliant infrastructure growths were noticed in safety-related troubles during the world. So, with multiplied call for for Security, Video-based Surveillance has grown to be an essential location for the research. An Intelligent Video Surveillance system essentially censored the performance, happenings, or converting data normally in terms of humans, cars or every other item from a distance by way of some electronic equipment (normally virtual digital camera). The scopes like prevention, detection, and intervention that have brought about the improvement of real and constant video surveillance structures can shrewd video processing abilities. In wide phrases, superior video-based surveillance could be defined as a shrewd video processing approach designed to help protect personnel by imparting reliable real-time alerts and to support green video evaluation for forensic investigations. This chapter offers the diverse requirements for designing a robust and reliable video surveillance machine. Also, it is mentioned the one-ofa-kind kinds of cameras required in one of a kind environmental conditions together with indoor and out of doors surveillance. Different modeling schemes are required for designing green surveillance machines under numerous illumination conditions.

Classification of Traffic Vehicle Density Using Deep Learning

The volume density of vehicles is a problem that often occurs in every city, as for the impact of vehicle density is congestion. Classification of vehicle density levels on certain roads is required because there are at least 7 vehicle density level conditions. Monitoring conducted by the police, the Department of Transportation and the organizers of the road currently using video-based surveillance such as CCTV that is still monitored by people manually. Deep Learning is an approach of synthetic neural network-based learning machines that are actively developed and researched lately because it has succeeded in delivering good results in solving various soft-computing problems, This research uses the convolutional neural network architecture. This research tries to change the supporting parameters on the convolutional neural network to further calibrate the maximum accuracy. After the experiment changed the parameters, the classification model was tested using K-fold cross-validation, confusion matrix and model exam with data testing. On the K-fold cross-validation test with an average yield of 92.83% with a value of K (fold) = 5, model testing is done by entering data testing amounting to 100 data, the model can predict or classify correctly i.e. 81 data.

Hand hygiene compliance monitoring: Do video-based technologies offer opportunities for the future?

Hand hygiene is universally recognised as the primary measure to reduce healthcare-associated infections. Studies have convincingly demonstrated a link between increased hand hygiene compliance and reductions in rates of healthcare-associated infections. Direct observation is considered the gold standard method for monitoring hand hygiene compliance. Despite the acknowledged benefits of this approach, recent literature has highlighted a range of issues impacting on the reliability and validity of this data collection technique. The rise of technology in healthcare provides opportunity for alternative methods that promise advantages over direct human observation. There have been no published examples of systems that are able to capture data consistent with all the WHO '5 Moments for Hand Hygiene'. In this paper we explore current human-based auditing practises for monitoring hand hygiene compliance and raise for discussion and debate video-based technologies to monitor hand hygiene compliance. We raise questions regarding hybrid approaches that incorporate both direct human observation and indirect video-based surveillance, and the possible advantages and disadvantages therein for monitoring hand hygiene compliance. We suggest that such methods have the potential to ameliorate, or minimise, the inherent biases associated with direct observation, notably the Hawthorne Effect. Future research into the utility of a hybrid approach to auditing, including the technical specifications, efficacy, cost effectiveness and acceptability of such a model is warranted.

Online learning of contexts for detecting suspicious behaviors in surveillance videos

One of the main incentives for implementing video-based surveillance systems is the urban security. During the last years, several approaches for automatic detection of suspicious events have been proposed. Those methods usually require a training stage before starting their operation. This means that previous to run time a representative dataset of interest events, that may occur in the future, must be available. Nevertheless, most real surveillance systems lack of that information, so many of those proposals results impractical.In this paper, a context online learning scheme for detecting suspicious behaviors on surveillance videos is proposed. Contextual information, which is inferred from videos of people in a scenario, allows detecting suspicious behaviors before an eventual criminal's final attack occur. The main attribute of the proposed approach is the capacity to start up its operation with a reduced training dataset. By an incremental learning process, which uses new data obtained during the online operation, the proposed scheme improves the performance over time achieving a better adaptation to conditions of each scenario.The proposed scheme was validated on two datasets. The first of them includes threats against a parked truck and its driver. The second testing dataset is composed of night assault scenes recorded in an urban environment. The experimental results demonstrate that the proposed method is able to learn incrementally from a reduced initial dataset, achieving a performance similar to batch-type systems trained with all data simultaneously and outperforming five state-of-the-art algorithms over violence detection.

A simple low-cost video-based surveillance system for a flash flood warning system

Flash flood is the most common and frequent natural disasters worldwide and can cause a loss of life and damage to property and the environment. In 2014, Manado city has experienced the most devastating flood that caused 19 people died, 10,884 homes destroyed, 2,091 people became homeless and estimated an economic loss of US$ 124 million. The huge loss was caused by the people and government did not receive information about water level rising on the watershed of the Tondano River. In conventional, decision maker needs the precipitation data and critical threshold value to issue the evacuation. There is an urgent requirement for end-to-end solution where the flash flood early warning system (FEWS) which can processed data in real-time can be integrated with Video-Based Surveillance System to take a role in saving peoples life. The aim of this study was to create a stand-alone system (Sensor Node) that could be placed on the watershed to monitor the water level rising using a microcomputer, a camera and a distance sensor. The aim of the system proposed was to measure the water level rising and when the threshold value was reached, the system would capture and display a real-time video. Future research can be done to “go-live” and posted the captured real-time video on the YouTube or Facebook in which people used them widely as a social media

Locally Statistical Dual-Mode Background Subtraction Approach

Due to the variety of background model in the real world, detecting changes in a video cannot be addressed exhaustively by a simple background subtraction method, especially with several motion detection challenges, such as dynamic background, camera jitter, intermittent object motion, and so on. In this paper, we propose an efficient background subtraction method, namely locally statistical dual-mode (LSD), for detecting moving objects in video-based surveillance systems. The method includes a local intensity pattern comparison algorithm for foreground segmentation by analyzing the homogeneity of intensity patterns of the input frame and the background model, in which the homogeneity is calculated by the mean and standard deviation of pixel intensity. Besides that, a dual-mode scheme is developed to temporally update the background model for the short- and long-term scenarios corresponding to sudden and gradual changes in the background. The advantage of this scheme is the allowance of updating the model in both pixel- and frame-wise manners simultaneously. The parameters used in both the local intensity pattern comparison algorithm and the dual-mode background model updating scheme are estimated for every input frame consecutively based on local and global statistical information of segmentation result. In experiments, the proposed LSD method is extensively evaluated on the Wallflower and CDnet2014 datasets; and remarkable performance demonstrates its preeminence to the many state-of-the-art background subtraction approaches in terms of segmentation accuracy and computational complexity.

Content-Aware Focal Plane Selection and Proposals for Object Tracking on Plenoptic Image Sequences.

Object tracking is a fundamental problem in computer vision since it is required in many practical applications including video-based surveillance and autonomous vehicles. One of the most challenging scenarios in the problem is when the target object is partially or even fully occluded by other objects. In such cases, most of existing trackers can fail in their task while the object is invisible. Recently, a few techniques have been proposed to tackle the occlusion problem by performing the tracking on plenoptic image sequences. Although they have shown promising results based on the refocusing capability of plenoptic images, there is still room for improvement. In this paper, we propose a novel focus index selection algorithm to identify an optimal focal plane where the tracking should be performed. To determine an optimal focus index, we use a focus measure to find maximally focused plane and a visual similarity to capture the plane where the target object is visible, and its appearance is distinguishably clear. We further use the selected focus index to generate proposals. Since the optimal focus index allows us to estimate the distance between the camera and the target object, we can more accurately guess the scale changes of the object in the image plane. Our proposal algorithm also takes the trajectory of the target object into account. We extensively evaluate our proposed techniques on three plenoptic image sequences by comparing them against the prior tracking methods specialized to the plenoptic image sequences. In experiments, our method provides higher accuracy and robustness over the prior art, and those results confirm that the merits of our proposed algorithms.