Stationary Video Camera Research Articles

Screen Use at Bedtime and Sleep Duration and Quality Among Youths

Although questionnaire-based cross-sectional research suggests that screen time before bed correlates with poor sleep, self-reported data seem unlikely to capture the complexity of modern screen use, requiring objective night-by-night measures to advance this field. To examine whether evening screen time is associated with sleep duration and quality that night in youths. This repeated-measures cohort study was performed from March to December 2021 in participant homes in Dunedin, New Zealand. Participants included healthy youths aged 11 to 14.9 years. Data were analyzed from October to November 2023. Objectively measured screen time, captured using wearable or stationary video cameras from 2 hours before bedtime until the first time the youth attempted sleep (shut-eye time) over 4 nonconsecutive nights. Video data were coded using a reliable protocol (κ = 0.92) to quantify device (8 options [eg, smartphone]) and activity (10 options [eg, social media]) type. Sleep duration and quality were measured objectively via wrist-worn accelerometers. The association of screen use with sleep measures was analyzed on a night-by-night basis using mixed-effects regression models including participant as a random effect and adjusted for weekends. Of the 79 participants (47 [59.5%] male; mean [SD] age, 12.9 [1.1] years), all but 1 had screen time before bed. Screen use in the 2 hours before bed had no association with most measures of sleep health that night (eg, mean difference in total sleep time, 0 minutes [95% CI, -3 to 20 minutes] for every 10 minutes more total screen time). All types of screen time were associated with delayed sleep onset but particularly interactive screen use (mean difference, 10 minutes; 95% CI, 4 to 16 minutes for every additional 10 minutes of interactive screen time). Every 10 minutes of additional screen time in bed was associated with shorter total sleep time (mean difference, -3 minutes; 95% CI, -6 to -1 minute). The mean difference in total sleep time was -9 minutes (95% CI, -16 to -2 minutes) for every 10 minutes of interactive screen use and -4 minutes (95% CI, -7 to 0 minutes) for passive screen use. In particular, gaming (mean difference, -17 minutes; 95% CI, -28 to -7 minutes for every 10 minutes of gaming) and multitasking (mean difference, -35 minutes; 95% CI, -67 to -4 minutes on nights with vs without multitasking) were associated with less total sleep time. In this repeated-measures cohort study, use of an objective method showed that screen time once in bed was associated with impairment of sleep, especially when screen time was interactive or involved multitasking. These findings suggest that current sleep hygiene recommendations to restrict all screen time before bed seem neither achievable nor appropriate.

Importance of complex sponges as habitat and feeding substrata for coral reef fishes

Coral reef fishes are usually assumed to be most strongly associated with reef-building corals. However, sponges can be a significant structural component of coral reef ecosystems and their framework can enhance the local abundance and biodiversity of fish assemblages. Little is known regarding the range of fish species using complex sponges as either shelter or feeding substrata. Here we use a combination of stationary video cameras and focal animal sampling to document fish species positively associated with complex sponges in Kimbe Bay, Papua New Guinea. Stationary cameras identified 45 fish species using the sponges for either shelter, feeding substrata or as sites for ambush predation. A guild of 10 individual fish species from five families (Blenniidae, Chaetodontidae, Gobiidae, Labridae and Pomacentridae) were observed to quantify sponge and other habitat use and compared with habitat availability to determine the level of sponge selectivity. One species, Pleurosicya elongata (the Slender Spongegoby), lived in obligate association with Ianthella basta (Elephant Ear sponge), and there was a positive relationship between sponge size and number of resident fish, however this was not significant for all life stages. Five other fish species appeared to preferentially select sponges as habitat (Amblyglyphidodon aureus, Chaetodon kleinii, Coradion chrysozonus, Escenius prooculis and Pomacentrus nigromanus), while for others, sponge use appeared incidental. When selectivity indices were calculated for specific sponge species it was apparent that some fishes exhibited preferences for particular sponge species or growth forms. These results suggest more fish species may be reliant on sponges than is widely appreciated.

REAL-TIME SMOKE DETECTION IN VIDEO

The paper considers algorithmic and software for early fire detection based on smoke detection from video sequences
 generated by a static video camera. To detect areas with smoke, an algorithm has been developed that allows
 you to select such areas on video frames that are characterized by a number of features: the presence of a stable
 directional movement, compliance with the color characteristics of smoke, and a decrease in the energy value of highfrequency
 components relative to the background model. The feature of the algorithm is a step-by-step spatiotemporal
 analysis of candidate areas, which provides satisfactory computational costs and real-time operation
 on modern computing tools for high-resolution video frames. The algorithm is implemented using the functions
 of the OpenCV computer vision library and multi-threaded processing. The features and main functionality of the software
 implemented as a stationary application are given. The results of experimental studies on the evaluation
 of the efficiency of the algorithm and its speed are presented.

A New Method for Validating and Generating Vehicle Trajectories From Stationary Video Cameras

Image processing based vehicle tracking is a powerful tool for monitoring traffic, but it is error prone. Relatively small errors that impede measuring time-series speed and acceleration can be hard to detect, e.g., 1 m positioning error in a 100 m long trajectory. This paper presents an efficient approach to separate the positioning errors from vehicle travel for evaluating image processing based vehicle trajectories. The approach starts with a spatiotemporal slice, STS, which is effectively a visual time-space diagram sampled from the video. This work skews the STS to flatten a given trajectory, eliminating the vehicle travel recorded in the trajectory. Positioning errors that were imperceptible relative to the distance traveled become readily apparent in the flattened track. Thus, providing a means to quickly assess reported trajectories from almost any image processing system against the true vehicle positions in the original video data. Recognizing that the flattening process works both ways, if errors are evident in a given trajectory, the STS method can also be used to quickly fix them. Thereby providing a path to accurate instantaneous speed and acceleration throughout the given trajectory. Alternatively, one can use this process to generate vehicle trajectories directly from the STS. While the main focus is longitudinal tracking, the process can also be used to assess (extract) the lateral position of a given vehicle. The method is evaluated using the NGSIM, Cityflow and UA-DETRAC datasets, in each case it is shown how this work can increase the fidelity of the given dataset.

Многофункциональная система безопасности угольных шахт – визуализация событий (горнотехнических процессов) с рабочего места шахтера

The paper reviews the use of mobile video monitoring equipment in coal mines. The most common option is the use of stationary video cameras with real-time video streaming to the mine dispatcher's control monitor via cables. Despite all the benefits of the information obtained, this method has certain limitations due to the specific features of the mine atmosphere, i.e. high humidity and dust levels, as well as the impossibility to organize video monitoring over the entire length of the mine workings. Therefore, mobile video monitoring equipment, both portable and vehicle-based, is efficient supplement to the stationary video cameras. The portable devices include smart phones and the battery-powered head lights with an integrated video camera, which have recently become very popular. In both cases, an important consideration, in addition to the actual video capturing, is the issue of transmitting video data to the top level, i.e. to the mine dispatcher's control panel. The following options are possible: connection to the mine wireless network hotspots via radio channel, reading the information in the lamp rooms when leaving the mine and real-time broadcasting from the mine to the top level. The assumption is made that in order to implement the fastest (and the most efficient) way that works without delays between capturing and transmitting of video data to the daylight surface, such as the latter of the options above, a communications infrastructure based on wireless and cable networks needs to be deployed in the mine workings. The required infrastructure is present in a number of systems designed to locate miners inside the mine workings as part of a multifunctional security system, which enables continuous radio communication of individual devices with infrastructure nodes and, therefore, real-time video data transmission.

Frame‐by‐frame annotation of video recordings using deep neural networks

AbstractVideo data are widely collected in ecological studies, but manual annotation is a challenging and time‐consuming task, and has become a bottleneck for scientific research. Classification models based on convolutional neural networks (CNNs) have proved successful in annotating images, but few applications have extended these to video classification. We demonstrate an approach that combines a standard CNN summarizing each video frame with a recurrent neural network (RNN) that models the temporal component of video. The approach is illustrated using two datasets: one collected by static video cameras detecting seal activity inside coastal salmon nets and another collected by animal‐borne cameras deployed on African penguins, used to classify behavior. The combined RNN‐CNN led to a relative improvement in test set classification accuracy over an image‐only model of 25% for penguins (80% to 85%), and substantially improved classification precision or recall for four of six behavior classes (12–17%). Image‐only and video models classified seal activity with very similar accuracy (88 and 89%), and no seal visits were missed entirely by either model. Temporal patterns related to movement provide valuable information about animal behavior, and classifiers benefit from including these explicitly. We recommend the inclusion of temporal information whenever manual inspection suggests that movement is predictive of class membership.

РЕАЛИЗАЦИЯ ОСНОВНЫХ НАПРАВЛЕНИЙ ПРОФИЛАКТИКИ ГРУППОВЫХ ПРАВОНАРУШЕНИЙ НЕСОВЕРШЕННОЛЕТНИХ ОСУЖДЕННЫХ В ВОСПИТАТЕЛЬНЫХ КОЛОНИЯХ

The federal legislation views the prevention of juvenile delinquency and antisocial actions as very significant, since an offense is understood as an administrative offense and a criminal offense, and antisocial actions include other offenses. This issue is topical for young offender institutions; however, the special category of detainees implies a special category of crimes and violations of the established order of serving the sentence (disciplinary responsibility of convicts). By the end of 2020, there were 949 people in 18 young offender institutions. According to the statistics, the convicts did not commit crimes, yet there were more than 545 other violations. In the current circumstances, it is necessary to improve the prevention of offenses by focusing on specific activities in penal institutions, including active use of non-contact supervision over the behavior of convicts (stationary video cameras, portable video recorders, biometric identification means, access control and management systems). The researchers and practitioners are discussing the possibility of using electronic monitoring systems to supervise the behavior of convicts. This will allow a better quality of supervision over the special contingent and more effective preventive activities in young offender institutions.

Study of the efficiency of detecting and recognizing drone images from a video stream

The authors have developed and experimentally tested an algorithm for processing a video stream of a stationary video camera. It consists of the stages of detecting moving objects and classifying these objects using a neural network. To detect moving objects, the methods of identifying moving objects against a stationary background and analyzing the history of motion were used. Based on the experimental data, the effectiveness of using the models of the background images of MOG, MOG2, KNN, GMG, CNT, GSOC, LSBP for solving the problem was analyzed. Recommendations for the choice of the parameters of these models were formulated. The selection criteria were as follows: high performance and low noise. Models of fully connected and convolutional neural networks were created and trained making it possible to classify 12 types of moving objects. Sets of images were created to train neural networks: drones, fragments of tree foliage, grass, clouds and insects. Based on the results of training and testing networks, recommendations are given for the number of network layers, the number of neurons in a layer, the number of convolutions to achieve maximum performance and recognition accuracy. Comparative analysis of the accuracy of drone classification using fully connected and convolutional networks when processing experimental data has proven the effectiveness of using convolutional networks. The dependence of the drone detection accuracy on the image size and, accordingly, on the distance to this drone is plotted.

Dynamic Spatial Predicted Background.

We present a novel method for online background modeling for static video cameras - Dynamic Spatial Predicted Background (DSPB). Our unique method employs a small subset of image pixels to predict the whole scene by exploiting pixel correlations (distant and close). DSPB acts as a hybrid model combining successful elements taken from two major approaches: local-adaptive that propose to fit a distribution pixelwise, and global-linear that reconstruct the background by finding a lowrank version of the scene. To our knowledge, this is the first attempt to combine these approaches in a unified system. DSPB models the scene as a superposition of illumination effects and predicts each pixel's value by a linear estimator comprised of only 5 pixels of the scene and can initialize the background starting from the 5th frame. By doing so, we keep the computational load low, allowing our method to be used in many real-time applications using simple hardware. The suggested prediction model of scene appearance is novel, and the scheme is very accurate and efficient computationally. We show the method merits on an application for video FG-BG separation, and how some of the main existing approaches may be challenged and how their drawbacks are less dominant in our model. Experimental results validate our findings, by computation speed and mean F-measure values on several public datasets. We also examine how results may improve by analyzing each video individually according to its content. DSPB can be successfully incorporated in other image processing tasks like change detection, video compression and video inpainting.

Multi-Frame Based Homography Estimation for Video Stitching in Static Camera Environments.

In this paper, a multi-frame based homography estimation method is proposed for video stitching in static camera environments. A homography that is robust against spatio-temporally induced noise can be estimated by intervals, using feature points extracted during a predetermined time interval. The feature point with the largest blob response in each quantized location bin, a representative feature point, is used for matching a pair of video sequences. After matching representative feature points from each camera, the homography for the interval is estimated by random sample consensus (RANSAC) on the matched representative feature points, with their chances of being sampled proportional to their numbers of occurrences in the interval. The performance of the proposed method is compared with that of the per-frame method by investigating alignment distortion and stitching scores for daytime and noisy video sequence pairs. It is shown that alignment distortion in overlapping regions is reduced and the stitching score is improved by the proposed method. The proposed method can be used for panoramic video stitching with static video cameras and for panoramic image stitching with less alignment distortion.

Autonomous Vehicular Surveillance using License Plate Recognition over Cloud Computing Architecture

License Plate Recognition (LPR) is the extraction and identification of licence plate numbers from license plates. The extraction process requires ample image pre-processing using normalization, gray scaling and edge removal techniques. These extracted plates can then be identified using image processing techniques such as neural networks and support vector machines. These license plates are captured using stationary video cameras, which extracts images from their feed as inputs into the image extraction algorithm. For the purposes of vehicular surveillance, these cameras are inefficient, as a lot of them will be required to monitor vehicles effectively. Hence there is a need for a larger scale model to carry out effective vehicular surveillance. For this purpose, the cameras embedded in self driving cars are utilised as replacements to stationary video cameras. These cameras have to advantage of being constantly mobile, hence being able to carry out a larger scale of surveillance. These cameras capture meaningful images of license plates from their video feed, and upload these images to the cloud using a Vehicular Cloud Computing (VCC) architecture. This centralized cloud carries out the image extraction and image processing tasks. The identified license plates can be used to monitor the cars they belong to. The cloud compares them to a database of license plates that are flagged by law enforcement. If the license plate is found to be flagged, then the respective law enforcement authorities are notified of the location of the car. If the plate belongs to a car with a history of misbehavior, the car capturing the plate is informed of thus, making it easier to safely navigate around the problematic driver.

Teacher's visual attention when scaffolding collaborative mathematical problem solving

Teachers' role in scaffolding students' problem-solving process is crucial. New technology provides researchers with possibilities to explore this aspect of teaching from the viewpoint of teacher attention. The aim of this mixed-method case study was to investigate the relation between a teacher's scaffolding intentions and his gaze behavior. The data was collected during a mathematics lesson using mobile gaze tracking devices, stationary video cameras, and interview. The results show that the teacher's scaffolding intentions affected his gaze targets significantly and that mobile gaze tracking can provide novel insight to situational processes of teacher-student interaction.

Reconstruction of road accident using video recording

Presented was the procedure of reconstructing a motorcycle-motor car collision with both vehicles moving in the same direction and performing simultaneous manoeuvres: a left turn and overtaking. Because of the conflicting reports of witnesses, the procedure also involved recordings from a stationary video camera. These data allowed us to retrace the movement and manoeuvres of vehicles as a function of time and then to identity the perpetrator of the accident.

Application of non-destructive methods for assessing rock pool fish assemblages on Lord Howe Island, Australia

Rock pool fish assemblages can be highly diverse, though are poorly studied in many locations. Where rock pool fishes have been studied, sampling has often been undertaken using destructive techniques, causing fish mortality and damage to pool ecosystems. There is, therefore, a need for greater understanding of non-destructive methods for evaluating rock pool fish assemblages. To improve knowledge in this area, we tested three non-destructive techniques: mini baited remote underwater videos (mini-BRUVs) which utilised a stationary video camera to record fish, visual censuses (VCs) where a roving observer recorded fish using a slate and observer operated videos (OOVs) where fish were recorded by an observer using a video camera.These methods were tested for their effectiveness in assessing fish assemblages, using data from rock pools on Lord Howe Island (LHI), in New South Wales (NSW), Australia. VCs and OOVs required significantly less total survey time than mini-BRUVs, whereas mini-BRUVs provided advantages in the detection of fishes in small pools (<25 m3). Examination of mini-BRUVs’ data, for 30 min samples, identified that the vast majority of fish species were detected within 15 min and indicated that a shorter 15 min sampling period would generally be suitable for collecting fish data from rock pools. The base-line data collected on fishes in LHI rock pools demonstrated the uniqueness of LHI assemblages, which were significantly different from those on the NSW coast. This study demonstrates that non-destructive sampling methods can be used in rock pool environments, making these methods suitable for sampling in inter-tidal habitats, especially within marine protected areas.

Competition Load Described by Objective and Subjective Methods During a Surfing Championship.

Fernández-Gamboa, I, Yanci, J, Granados, C, Freemyer, B, and Cámara, J. Competition load described by objective and subjective methods during a surfing championship. J Strength Cond Res 32(5): 1329-1335, 2018-The aims of this study were to describe the competition load of surfers during a single heat through objective and subjective methods and to analyze the relationship between objective and subjective methods with the judges' score. Ten competitive surfers were fitted with a global positioning system (GPS) during a competitive heat. The GPS was synchronized with a chronometer and a stationary video camera to identify the surfer's specific actions. After the end of each heat, participants were assessed for the rating of perceived respiratory and muscular exertion (RPEres, RPEmus), and also, official scores from every participant were collected. A very large significant relationship between wave-riding distance and respiratory perceived exertion heat load (RPEres HL, r = 0.79; ±0.26 confidence limit [CL], p < 0.01, 99.5/0.4/0.1, very likely) was found. Active time was also very large and significantly related to both RPEres HL (r = 0.75; ±0.29 CL, p < 0.05, 99.0/0.8/0.2, very likely) and muscular perceived exertion heat load (RPEmus HL, r = 0.83; ±0.22 CL, p < 0.01, 99.8/0.2/0.0, most likely). Very large significant correlation was obtained between the RPEres and score (r = 0.83; ±0.22 CL, p < 0.01, 99.8/0.2/0.0, most likely). The subjective method seems to be a good instrument to assess the HL of a surf competition. Wave characteristics seem to be an important factor in perceived exertion during competitive surfing.

Influence of predation risk on the sheltering behaviour of the coral-dwelling damselfish, Pomacentrus moluccensis

Predation is a key ecosystem function, especially in high diversity systems such as coral reefs. Not only is predation one of the strongest top-down controls of prey population density, but it also is a strong driver of prey behaviour and function through non-lethal effects. We ask whether predation risk influences sheltering behaviour of damselfish living in mutualism with branching corals. Host corals gain multiple advantages from the mutualistic relationship which are determined by the strength of damselfish sheltering. Distance travelled by the Lemon Damselfish Pomacentrus moluccensis away from their host colony was measured here as a proxy for sheltering strength and was expected to be shortest under highest predation risk. Predation risk, defined as a function of predator abundance and activity, turbidity and habitat complexity, was quantified at four reef slope sites in Kepulauan Seribu, Indonesia. Damselfish sheltering strength was measured using stationary unmanned video cameras. Small damselfish (< 2 cm) increased their sheltering strength under high turbidity. Predator feeding activity, but not abundance, influenced damselfish sheltering strength. Contrary to our expectations, sheltering behaviour of adult damselfish decreased under high predator activity. While these observations are in line with risk-averse behaviour by juvenile P. moluccensis, they may indicate the presence of sentinel behaviour in the adults of this species. Habitat complexity seemed to be less important as a driver of damselfish behaviour. These counterintuitive results may indicate complex social behaviour and age-specific strategies for predator avoidance.

Google Glass Video Capture of Cardiopulmonary Resuscitation Events: A Pilot Simulation Study

Video recording of resuscitation from fixed camera locations has been used to assess adherence to guidelines and provide feedback on performance. However, inpatient cardiac arrests often happen in unpredictable locations and crowded rooms, making video recording of these events problematic. We sought to understand the feasibility of Google Glass (GG) as a method for recording inpatient cardiac arrests and capturing salient resuscitation factors for post-event review. This observational study involved recording simulated cardiac arrest events on inpatient medical wards. Each simulation was reviewed by 3 methods: in-room physician direct observation, stationary video camera (SVC), and GG. Nurse and physician specialists analyzed the videos for global visibility and audibility, as well as recording quality of predefined resuscitation events and behaviors. Resident code leaders were surveyed regarding attitudes toward GG use in the clinical emergency setting. Of 11 simulated cardiac arrest events, 9 were successfully recorded by all observation methods (1 GG failure, 1 SVC failure). GG was judged slightly better than SVC recording for average global visualization (3.95 versus 3.15, P = .0003) and average global audibility (4.77 versus 4.42, P = .002). Of the GG videos, 19% had limitations in overall interpretability compared with 35% of SVC recordings (P = .039). All 10 survey respondents agreed that GG was easy to use; however, 2 found it distracting and 3 were uncomfortable with future use during actual resuscitations. GG is a feasible and acceptable method for capturing simulated resuscitation events in the inpatient setting.

Quantifying the High-Speed Running and Sprinting Profiles of Elite Female Soccer Players During Competitive Matches Using an Optical Player Tracking System.

The aim of this study was to determine the high-speed running and sprinting profiles of elite female soccer players during competitive matches using a new Optical Player Tracking system. Eight stationary video cameras were positioned at vantage points surrounding the soccer field so that when each camera view was combined, the entire field could be viewed simultaneously. After each match, an optical player tracking system detected the coordinates (x, y) of each player for every video frame. Algorithms applied to the x and y coordinates were used to determine activity variables for 12 elite female players across 7 competitive matches. Players covered 9,220-10,581 m of total distance, 1,772-2,917 m of high-speed running (3.4-5.3 m·s) distance, and 417-850 m of sprinting (>5.4 m·s) distance, with variations between positional groups (p < 0.001; partial η = 0.444-0.488). Similarly, the number of high-speed runs differed between positional groups (p = 0.002; partial η = 0.342), and a large proportion of high-speed runs (81-84%) and sprints (71-78%) were performed over distances less than 10 m. Mean time between high-speed runs (13.9 ± 4.4 seconds) and sprints (86.5 ± 38.0 seconds) varied according to playing position (p < 0.001; partial η = 0.409) and time of the match (p < 0.001; partial η = 0.113-0.310). The results of this study can be used to design match-specific conditioning drills and shows that coaches should take an individualized approach to training load monitoring according to position.

Resident Performance Assessment: A Crucial Area for Development

Plastic surgery resident assessment has become the next element in resident education that must be developed. Reliable, comprehensive resident evaluation is currently difficult to achieve because of inconsistent assessor participation and an incomplete assessment toolbox. Recent advances in technology may be able to help, but they require judicious application. In particular, attention should be paid to assessing all levels of competence. Miller’s pyramid of assessment can help guide the process.1 The foundation of the pyramid is assessment of what a resident “knows.” Multiple choice or short answer examinations are a familiar tool that evaluates a resident’s medical knowledge. Smart phone applications and the online Plastic Surgery Education Network are an improvement over the older paper exams in which they provide both content and continuous assessment of knowledge and immediate feedback. Residents are able to test themselves at their own pace whereas teachers can monitor their progress and provide tailored education based on demonstrated need. The next level up on the pyramid is assessing what a resident “knows how” to do. Oral examinations do a good job of eliciting a plastic surgery resident’s knowledge of how to deliver good patient care or operate. The cognitive aspects of surgery can also be tested by computer simulators, drawing, and scripting. A resident’s ability to draw a procedure correlates well with the ability to perform the procedure. A resident’s operative script gives insight into the ability to plan and adapt, and the script serves as a permanent technical manual for the resident to revise over the course of a career. Overall, plastic surgery has adequate tools to assess the first 2 levels of the pyramid. The challenge lies in assessing the next 2 levels, particularly in the operating room. The third level of the assessment pyramid has the resident “show how.” In the clinic, standardized patients or simulated scenarios provide a safe way of allowing a resident to demonstrate a required competence before engaging with a real patient. For the operating room, surgical skills labs with cadavers or animal models and both low and high fidelity surgical simulators are excellent assessment tools. Particularly important is the ability of these tools to both teach and provide meaningful feedback in a way that does not put a patient at risk. The top of Miller’s pyramid is “does,” assessing a resident treating a real patient in the clinic or operating room. Portable sport cameras, stationary video cameras, and in-line recording of microsurgery allow direct, valid assessment of a resident’s ability to manage a patient or perform a surgical technique when filtered through the Objective Structured Clinical Examination or Objective Structured Assessment of Technical Skills rubrics available on the American Council for Graduate Medical Education website. Smart phone applications that allow for immediate intake and recording of an operative assessment can increase assessor participation, but in the end, technology cannot eliminate the need for an experienced surgeon to be the ultimate resident assessor.

Using an unbaited stationary video system to investigate the behaviour and interactions of bull sharks Carcharhinus leucas under an aquaculture farm

Bull sharks Carcharhinus leucas are common along the coast of Reunion Island (South-West Indian Ocean) and were suspected to aggregate in the vicinity of an aquaculture farm in Saint-Paul Bay on the west coast. In order to understand the behaviour and interaction of bull sharks near aquaculture cages at Saint-Paul Bay, we deployed an experimental unbaited stationary video camera. From 175 hours of recording during daylight hours from March to April 2012, eight individual female bull sharks (seven adults and one immature) were identified based on their natural markings. These sharks were resighted between 3 and 45 times. Residency analysis revealed site attachment under the aquaculture cages for at least three individuals over the course of the study. Recorded behaviours included intraspecific social interactions such as synchronised swimming. Social interactions and relatively strong paired associations for two pairs of females suggest some level of sociality among bull sharks around Reunion Island. Overall, our results demonstrate the utility of unbaited video systems to monitor the behaviour of adult coastal sharks.