Video analytics for the detection of near-miss incidents on approach to railway level crossings

Recent modelling of socio-economic costs by the Australian railway industry in 2010 has estimated the cost of level crossing accidents to exceed AU$116 million annually. To better understand causal factors that contribute to these accidents, the Cooperative Research Centre for Rail Innovation is running a project entitled Baseline Level Crossing Video. The project aims to improve the recording of level crossing safety data by developing an intelligent system capable of detecting near-miss incidents and capturing quantitative data around these incidents. To detect near-miss events at railway level crossings a video analytics module is being developed to analyse video footage obtained from forward-facing cameras installed on trains. This paper presents a vision base approach for the detection of these near-miss events. The video analytics module is comprised of object detectors and a rail detection algorithm, allowing the distance between a detected object and the rail to be determined. An existing publicly available Histograms of Oriented Gradients (HOG) based object detector algorithm is used to detect various types of vehicles in each video frame. As vehicles are usually seen from a sideway view from the cabin’s perspective, the results of the vehicle detector are verified using an algorithm that can detect the wheels of each detected vehicle. Rail detection is facilitated using a projective transformation of the video, such that the forward-facing view becomes a bird’s eye view. Line Segment Detector is employed as the feature extractor and a sliding window approach is developed to track a pair of rails. Localisation of the vehicles is done by projecting the results of the vehicle and rail detectors on the ground plane allowing the distance between the vehicle and rail to be calculated. The resultant vehicle positions and distance are logged to a database for further analysis. We present preliminary results regarding the performance of a prototype video analytics module on a data set of videos containing more than 30 different railway level crossings. The video data is captured from a journey of a train that has passed through these level crossings.

Near-miss event detection at railway level crossings

Recent modelling of socio-economic costs by the Australian railway industry in 2010 has estimated the cost of level crossing accidents to exceed AU$116 million annually. To better understand the causal factors of these accidents, a video analytics application is being developed to automatically detect near-miss incidents using forward facing videos from trains. As near-miss events occur more frequently than collisions, by detecting these occurrences there will be more safety data available for analysis. The application that is being developed will improve the objectivity of near-miss reporting by providing quantitative data about the position of vehicles at level crossings through the automatic analysis of video footage. In this paper we present a novel method for detecting near-miss occurrences at railway level crossings from video data of trains. Our system detects and localizes vehicles at railway level crossings. It also detects the position of railways to calculate the distance of the detected vehicles to the railway centerline. The system logs the information about the position of the vehicles and railway centerline into a database for further analysis by the safety data recording and analysis system, to determine whether or not the event is a near-miss. We present preliminary results of our system on a dataset of videos taken from a train that passed through 14 railway level crossings. We demonstrate the robustness of our system by showing the results of our system on day and night videos.

Improving Near-Miss Event Detection Rate at Railway Level Crossings

Even though crashes between trains and road users are rare events at railway level crossings, they are one of the major safety concerns for the Australian railway industry. Nearmiss events at level crossings occur more frequently, and can provide more information about factors leading to level crossing incidents. In this paper we introduce a video analytic approach for automatically detecting and localizing vehicles from cameras mounted on trains for detecting near-miss events. To detect and localize vehicles at level crossings we extract patches from an image and classify each patch for detecting vehicles. We developed a region proposals algorithm for generating patches, and we use a Convolutional Neural Network (CNN) for classifying each patch. To localize vehicles in images we combine the patches that are classified as vehicles according to their CNN scores and positions. We compared our system with the Deformable Part Models (DPM) and Regions with CNN features (R-CNN) object detectors. Experimental results on a railway dataset show that the recall rate of our proposed system is 29% higher than what can be achieved with DPM or R-CNN detectors.

Brazilian network for the surveillance of maternal potentially life threatening morbidity and maternal near-miss and a multidimensional evaluation of their long term consequences

Background. It has been suggested that the study of women who survive life-threatening complications related to pregnancy (maternal near-miss cases) may represent a practical alternative to surveillance of maternal morbidity/mortality since the number of cases is higher and the woman herself is able to provide information on the difficulties she faced and the long-term repercussions of the event. These repercussions, which may include sexual dysfunction, postpartum depression and posttraumatic stress disorder, may persist for prolonged periods of time, affecting women's quality of life and resulting in adverse effects to them and their babies. Objective. The aims of the present study are to create a nationwide network of scientific cooperation to carry out surveillance and estimate the frequency of maternal near-miss cases, to perform a multicenter investigation into the quality of care for women with severe complications of pregnancy, and to carry out a multidimensional evaluation of these women up to six months. Methods/Design. This project has two components: a multicenter, cross-sectional study to be implemented in 27 referral obstetric units in different geographical regions of Brazil, and a concurrent cohort study of multidimensional analysis. Over 12 months, investigators will perform prospective surveillance to identify all maternal complications. The population of the cross-sectional component will consist of all women surviving potentially life-threatening conditions (severe maternal complications) or life-threatening conditions (the maternal near miss criteria) and maternal deaths according to the new WHO definition and criteria. Data analysis will be performed in case subgroups according to the moment of occurrence and determining cause. Frequencies of near-miss and other severe maternal morbidity and the association between organ dysfunction and maternal death will be estimated. A proportion of cases identified in the cross-sectional study will comprise the cohort of women for the multidimensional analysis. Various aspects of the lives of women surviving severe maternal complications will be evaluated 3 and 6 months after the event and compared to a group of women who suffered no severe complications in pregnancy. Previously validated questionnaires will be used in the interviews to assess reproductive function, posttraumatic stress, functional capacity, quality of life, sexual function, postpartum depression and infant development. © 2009 Cecatti et al.

Fatores determinantes da near miss materna em uma unidade de terapia intensiva obstétrica: avaliação das competências profissionais da área médica

Objective: Evaluate the determinants of morbidity and mortality in an obstetric intensive care unit and professional medical skills of students/residents at a university hospital. Methods: observational cross - sectional with 492 pregnant/pue rperal women and 261 students/residents. Patients were admitted to the obstetric intensive care unit during a year, being informed about the proposals of the study and a questionnaire was applied. The analysis was performed using Microsoft Excel 2013 and G raphPad6. Chi - square tests were used to evaluate risk factors and student t test evaluates resident/students' skills concerning the cognitive test and the Mini - Cex. Results: the main risk factors to near miss were: non - white race (OR = 2.527; RR = 2.342) ; marital status(married women) (OR = 7.968; RR = 7.113) , schooling (primary) (OR = 3.177 ; RR = 2.829) , from country town (OR = 4.643 ; RR = 4.087), low income (OR = 7014 ; RR = 5.554) , gestational hypertensive disorders (OR = 16.35 ; RR = 13.27) , re alization of pre - natal (OR = 5.023 ; RR = 4.254) and C - section before labor(OR = 39.21 ; RR = 31.25). In cognitive/Mini - cex analysis were noted significant difference in the performance of students on the subject (3.75 ± 0.93, 4.03 ± 0.94 and 4.88 ± 0.35). We still observed the best performance of residents, when compared to graduation students (p < 0.01). Conclusions: the prevalence of near miss was associated with socioeconomic/clinics factors and care issues, revealing the importance of interventions to improve these indicators. In addition, we suggest a better curriculum insertion of this subject in the medical Course disciplines due the importance to avoid the near miss through of adequacy of medical education.

Near miss materno e iniquidades em saúde: análise de determinantes contextuais no Rio Grande do Norte, Brasil

The scope of this study was to identify socioeconomic contextual and health care factors in primary care associated with maternal near misses and their marker conditions. This is an ecological study that used aggregated data of 63 clusters formed by the municipalities of State of Rio Grande do Norte, Brazil, using the Skater method of area regionalization, as the unit of analysis. The ratio of maternal near misses and their marker conditions were obtained from the Hospital Information System of the Brazilian Unified Health System. In multiple linear regression analysis, there was a significant association between maternal near misses and variables of poverty and poor primary health care. Hypertensive disorders were also associated with poverty and poor primary care and the occurrence of hemorrhaging was associated with infant mortality. It was observed that the occurrence of maternal near misses is linked to unfavorable socioeconomic conditions and poor quality health care that are a reflection of public policies that accentuate health inequalities.

Near miss materno e iniquidades em saúde: análise de determinantes contextuais no Rio Grande do Norte, Brasil

The scope of this study was to identify socioeconomic contextual and health care factors in primary care associated with maternal near misses and their marker conditions. This is an ecological study that used aggregated data of 63 clusters formed by the municipalities of State of Rio Grande do Norte, Brazil, using the Skater method of area regionalization, as the unit of analysis. The ratio of maternal near misses and their marker conditions were obtained from the Hospital Information System of the Brazilian Unified Health System. In multiple linear regression analysis, there was a significant association between maternal near misses and variables of poverty and poor primary health care. Hypertensive disorders were also associated with poverty and poor primary care and the occurrence of hemorrhaging was associated with infant mortality. It was observed that the occurrence of maternal near misses is linked to unfavorable socioeconomic conditions and poor quality health care that are a reflection of public policies that accentuate health inequalities.

Prevalência e fatores associados ao Near Miss Materno: inquérito populacional em uma capital do Nordeste Brasileiro

The scope of this study was to determine the prevalence of near misses and complications during pregnancy and the puerperal period, identifying the main clinical and intervention markers and socioeconomic and demographic factors associated with near misses. It involved a cross-sectional, population-based and probabilistic study with multi-stage complex sampling design conducted in Natal, State of Rio Grande do Norte, Brazil. A validated questionnaire was given to 848 women aged 15 to 49 identified in 8,227 households in 60 census sectors. In theanalysis of associations, the Chi-square test applied and calculated the prevalence ratio (PR) with Confidence Interval (CI) of 95% and 5% significance. The prevalence of maternal near misses was 41.1/1000LB, with hospitalization in an Intensive Care Unit (19.1/1000LB) and eclampsia (13.5/1000LB) being the most important markers. The prevalence of complications during pregnancy and the puerperal period was 21.2%. The highest prevalence of near misses was observed in older women, of black/brown race and low socioeconomic status. Conducting population surveys is feasible and may add important information to the study of near misses and the markers highlight the need for enhancing maternal care to reduce health inequality.

Prevalência e fatores associados ao Near Miss Materno: inquérito populacional em uma capital do Nordeste Brasileiro

The scope of this study was to determine the prevalence of near misses and complications during pregnancy and the puerperal period, identifying the main clinical and intervention markers and socioeconomic and demographic factors associated with near misses. It involved a cross-sectional, population-based and probabilistic study with multi-stage complex sampling design conducted in Natal, State of Rio Grande do Norte, Brazil. A validated questionnaire was given to 848 women aged 15 to 49 identified in 8,227 households in 60 census sectors. In theanalysis of associations, the Chi-square test applied and calculated the prevalence ratio (PR) with Confidence Interval (CI) of 95% and 5% significance. The prevalence of maternal near misses was 41.1/1000LB, with hospitalization in an Intensive Care Unit (19.1/1000LB) and eclampsia (13.5/1000LB) being the most important markers. The prevalence of complications during pregnancy and the puerperal period was 21.2%. The highest prevalence of near misses was observed in older women, of black/brown race and low socioeconomic status. Conducting population surveys is feasible and may add important information to the study of near misses and the markers highlight the need for enhancing maternal care to reduce health inequality.

Beyond maternal death: improving the quality of maternal care through national studies of ‘near-miss’ maternal morbidity

Knight M, Acosta C, Brocklehurst P, Cheshire A, Fitzpatrick K, Hinton L, Jokinen M, Kemp B, Kurinczuk JJ, Lewis G, Lindquist A, Locock L, Nair M, Patel N, Quigley M, Ridge D, Rivero-Arias O, Sellers S, Shah A on behalf of the UKNeS coapplicant group. Background Studies of maternal mortality have been shown to result in important improvements to women’s health. It is now recognised that in countries such as the UK, where maternal deaths are rare, the study of near-miss severe maternal morbidity provides additional information to aid disease prevention, treatment and service provision. Objectives To (1) estimate the incidence of specific near-miss morbidities; (2) assess the contribution of existing risk factors to incidence; (3) describe different interventions and their impact on outcomes and costs; (4) identify any groups in which outcomes differ; (5) investigate factors associated with maternal death; (6) compare an external confidential enquiry or a local review approach for investigating quality of care for affected women; and (7) assess the longer-term impacts. Methods Mixed quantitative and qualitative methods including primary national observational studies, database analyses, surveys and case studies overseen by a user advisory group. Setting Maternity units in all four countries of the UK. Participants Women with near-miss maternal morbidities, their partners and comparison women without severe morbidity. Main outcome measures The incidence, risk factors, management and outcomes of uterine rupture, placenta accreta, haemolysis, elevated liver enzymes and low platelets (HELLP) syndrome, severe sepsis, amniotic fluid embolism and pregnancy at advanced maternal age (≥ 48 years at completion of pregnancy); factors associated with progression from severe morbidity to death; associations between severe maternal morbidity and ethnicity and socioeconomic status; lessons for care identified by local and external review; economic evaluation of interventions for management of postpartum haemorrhage (PPH); women’s experiences of near-miss maternal morbidity; long-term outcomes; and models of maternity care commissioned through experience-led and standard approaches. Results Women and their partners reported long-term impacts of near-miss maternal morbidities on their physical and mental health. Older maternal age and caesarean delivery are associated with severe maternal morbidity in both current and future pregnancies. Antibiotic prescription for pregnant or postpartum women with suspected infection does not necessarily prevent progression to severe sepsis, which may be rapidly progressive. Delay in delivery, of up to 48 hours, may be safely undertaken in women with HELLP syndrome in whom there is no fetal compromise. Uterine compression sutures are a cost-effective second-line therapy for PPH. Medical comorbidities are associated with a fivefold increase in the odds of maternal death from direct pregnancy complications. External reviews identified more specific clinical messages for care than local reviews. Experience-led commissioning may be used as a way to commission maternity services. Limitations This programme used observational studies, some with limited sample size, and the possibility of uncontrolled confounding cannot be excluded. Conclusions Implementation of the findings of this research could prevent both future severe pregnancy complications as well as improving the outcome of pregnancy for women. One of the clearest findings relates to the population of women with other medical and mental health problems in pregnancy and their risk of severe morbidity. Further research into models of pre-pregnancy, pregnancy and postnatal care is clearly needed.

Beyond the numbers of maternal near-miss in Rwanda - a qualitative study on women's perspectives on access and experiences of care in early and late stage of pregnancy.

BACKGROUND: Rwanda has made remarkable progress in decreasing the number of maternal deaths, yet women still face morbidities and mortalities during pregnancy. We explored care-seeking and experiences of maternity care among women who suffered a near-miss event during either the early or late stage of pregnancy, and identified potential health system limitations or barriers to maternal survival in this setting. METHODS: A framework of Naturalistic Inquiry guided the study design and analysis, and the 'three delays' model facilitated data sorting. Participants included 47 women, who were interviewed at three hospitals in Kigali, and 14 of these were revisited in their homes, from March 2013 to April 2014. RESULTS: The women confronted various care-seeking barriers depending on whether the pregnancy was wanted, the gestational age, insurance coverage, and marital status. Poor communication between the women and healthcare providers seemed to result in inadequate or inappropriate treatment, leading some to seek either traditional medicine or care repeatedly at biomedical facilities. CONCLUSION: Improved service provision routines, information, and amendments to the insurance system are suggested to enhance prompt care-seeking. Additionally, we strongly recommend a health system that considers the needs of all pregnant women, especially those facing unintended pregnancies or complications in the early stages of pregnancy.

Efficacy of proxy definitions for identification of fatigue/sleep-related crashes : an Australian evaluation

Fatigue/sleepiness is recognised as an important contributory factor in fatal and serious injury road traffic incidents (RTIs), however, identifying fatigue/sleepiness as a causal factor remains an uncertain science. Within Australia attending police officers at a RTI report the causal factors; one option is fatigue/sleepiness. In some Australian jurisdictions police incident databases are subject to post hoc analysis using a proxy definition for fatigue/sleepiness. This secondary analysis identifies further RTIs caused by fatigue/sleepiness not initially identified by attending officers. The current study investigates the efficacy of such proxy definitions for attributing fatigue/sleepiness as a RTI causal factor. Over 1600 Australian drivers were surveyed regarding their experience and involvement in fatigue/sleep-related RTIs and near-misses during the past five years. Driving while fatigued/sleepy had been experienced by the majority of participants (66.0% of participants). Fatigue/sleep-related near misses were reported by 19.1% of participants, with 2.4% being involved in a fatigue/sleep-related RTI. Examination of the characteristics for the most recent event (either a near miss or crash) found that the largest proportion of incidents (28.0%) occurred when commuting to or from work, followed by social activities (25.1%), holiday travel (19.8%), or for work purposes (10.1%). The fatigue/sleep related RTI and near-miss experience of a representative sample of Australian drivers does not reflect the proxy definitions used for fatigue/sleepiness identification. In particular those RTIs that occur in urban areas and at slow speeds may not be identified. While important to have a strategy for identifying fatigue/sleepiness related RTIs proxy measures appear best suited to identifying specific subsets of such RTIs.

The Australian naturalistic driving study: From beginnings to launch

The Australian Naturalistic Driving Study (ANDS), a ground-breaking study of Australian driver behaviour and performance, was officially launched on April 21st, 2015 at UNSW. The ANDS project will provide a realistic perspective on the causes of vehicle crashes and near miss crash events, along with the roles speeding, distraction and other factors have on such events. A total of 360 volunteer drivers across NSW and Victoria - 180 in NSW and 180 in Victoria - will be monitored by a Data Acquisition System (DAS) recording continuously for 4 months their driving behaviour using a suite of cameras and sensors. Participants’ driving behaviour (e.g. gaze), the behaviour of their vehicle (e.g. speed, lane position) and the behaviour of other road users with whom they interact in normal and safety-critical situations will be recorded. Planning of the ANDS commenced over two years ago in June 2013 when the Multi-Institutional Agreement for a grant supporting the equipment purchase and assembly phase was signed by parties involved in this large scale $4 million study (5 university accident research centres, 3 government regulators, 2 third party insurers and 2 industry partners). The program’s second development phase commenced a year later in June 2014 after a second grant was awarded. This paper presents an insider's view into that two year process leading up to the launch, and outlines issues that arose in the set-up phase of the study and how these were addressed. This information will be useful to other organisations considering setting up an NDS.

Using cooperative vehicle systems to collect near collision incidents events : a simulation study

On the road, near collision events (also close calls or near-miss incidents) largely outnumber actual crashes, yet most of them can never be recorded by current traffic data collection technologies or crashes analysis tools. The analysis of near collisions data is an important step in the process of reducing the crash rate. There have been several studies that have investigated near collisions; to our knowledge, this is the first study that uses the functionalities provided by cooperative vehicles to collect near misses information. We use the VISSIM traffic simulator and a custom C++ engine to simulate cooperative vehicles and their ability to detect near collision events. Our results showed that, within a simple simulated environment, adequate information on near collision events can be collected using the functionalities of cooperative perception systems. The relationship between the ratio of detected events and the ratio of equipped vehicle was shown to closely follow a squared law, and the largest source of nondetection was packet loss instead of packet delays and GPS imprecision.

Building an augmented map for road risk assessment

Recent road safety statistics show that the decades-long fatalities decreasing trend is stopping and stagnating. Statistics further show that crashes are mostly driven by human error, compared to other factors such as environmental conditions and mechanical defects. Within human error, the dominant error source is perceptive errors, which represent about 50% of the total. The next two sources are interpretation and evaluation, which accounts together with perception for more than 75% of human error related crashes. Those statistics show that allowing drivers to perceive and understand their environment better, or supplement them when they are clearly at fault, is a solution to a good assessment of road risk, and, as a consequence, further decreasing fatalities. To answer this problem, currently deployed driving assistance systems combine more and more information from diverse sources (sensors) to enhance the driver's perception of their environment. However, because of inherent limitations in range and field of view, these systems' perception of their environment remains largely limited to a small interest zone around a single vehicle. Such limitations can be overcomed by increasing the interest zone through a cooperative process. Cooperative Systems (CS), a specific subset of Intelligent Transportation Systems (ITS), aim at compensating for local systems' limitations by associating embedded information technology and intervehicular communication technology (IVC). With CS, information sources are not limited to a single vehicle anymore. From this distribution arises the concept of extended or augmented perception. Augmented perception allows extending an actor's perceptive horizon beyond its "natural" limits not only by fusing information from multiple in-vehicle sensors but also information obtained from remote sensors. The end result of an augmented perception and data fusion chain is known as an augmented map. It is a repository where any relevant information about objects in the environment, and the environment itself, can be stored in a layered architecture. This thesis aims at demonstrating that augmented perception has better performance than noncooperative approaches, and that it can be used to successfully identify road risk. We found it was necessary to evaluate the performance of augmented perception, in order to obtain a better knowledge on their limitations. Indeed, while many promising results have already been obtained, the feasibility of building an augmented map from exchanged local perception information and, then, using this information beneficially for road users, has not been thoroughly assessed yet. The limitations of augmented perception, and underlying technologies, have not be thoroughly assessed yet. Most notably, many questions remain unanswered as to the IVC performance and their ability to deliver appropriate quality of service to support life-saving critical systems. This is especially true as the road environment is a complex, highly variable setting where many sources of imperfections and errors exist, not only limited to IVC. We provide at first a discussion on these limitations and a performance model built to incorporate them, created from empirical data collected on test tracks. Our results are more pessimistic than existing literature, suggesting IVC limitations have been underestimated. Then, we develop a new CS-applications simulation architecture. This architecture is used to obtain new results on the safety benefits of a cooperative safety application (EEBL), and then to support further study on augmented perception. At first, we confirm earlier results in terms of crashes numbers decrease, but raise doubts on benefits in terms of crashes' severity. In the next step, we implement an augmented perception architecture tasked with creating an augmented map. Our approach is aimed at providing a generalist architecture that can use many different types of sensors to create the map, and which is not limited to any specific application. The data association problem is tackled with an MHT approach based on the Belief Theory. Then, augmented and single-vehicle perceptions are compared in a reference driving scenario for risk assessment,taking into account the IVC limitations obtained earlier; we show their impact on the augmented map's performance. Our results show that augmented perception performs better than non-cooperative approaches, allowing to almost tripling the advance warning time before a crash. IVC limitations appear to have no significant effect on the previous performance, although this might be valid only for our specific scenario. Eventually, we propose a new approach using augmented perception to identify road risk through a surrogate: near-miss events. A CS-based approach is designed and validated to detect near-miss events, and then compared to a non-cooperative approach based on vehicles equiped with local sensors only. The cooperative approach shows a significant improvement in the number of events that can be detected, especially at the higher rates of system's deployment.