Black Box Camera Research Articles

A method for estimating the vehicle impact speed in collision with stationary vehicle or barrier using camera video record

In Korea, most drivers equip their vehicles with a black box camera, used to estimate the speed of the vehicle in case of traffic accidents such as car-to-car collisions, single-car accidents, and car-to-pedestrian accidents. Traditionally, speed estimation using the camera video has been done by using the distance between the road lane, surrounding structures, etc.However, when the following vehicle collides with a stationary leading vehicle or barrier, and no road lanes or structures are visible around the driving path in the black box camera video of the following vehicle, the speed of the following vehicle could be just estimated according the deformations on the collided vehicles or objects.Therefore, in this study, when a vehicle collision occurs, a virtual parallel line is drawn between the leading vehicle and the following vehicle in the black box camera image of the following vehicle and analyzed geometrically to estimate the distance and the speed of the following vehicle.After the vehicle collision experiment, the speed was estimated using the method proposed in this study using the captured black box video. As a result, it was confirmed that the accuracy was high when the collision speed was low. So The method proposed in this study is expected to be used to estimate collision speed in the cases of low impact speed in the rear end collisions.

Scenario establishment and characteristic analysis of intersection collision accidents for advanced driver assistance systems

Objective: Intersection collision is the most common type of vehicle accidents, and it is more complicated and has more variables compared to straight road collisions. Intersection collisions are also a major obstacle to practical application of self-driving technology. This calls for more studies and development of Intersection-Advanced Driver Assistance System (I-ADAS) which can be applied to various scenarios of intersection collisions.Method: In this study, NHTSA FARS and NASS-CDS DB from the period of 2013–2015 were used to analyze the circumstances and severity of damage in intersection collisions. With these analysis results, 17 possible vehicle collision scenarios were established based on the travel directions and relative positions of the two vehicles. The 17 accident scenarios were categorized into nine groups based on the travel direction and relative positions, and accident characteristics of each group, such as the severity of injuries, were analyzed.Results: Based on these characteristics, a method of qualitatively predicting and avoiding an intersection collision accident using a black box camera mounted on vehicles is introduced.Conclusion: When classifying the collision types, it was done with the consideration of making car-mounted camera-based intersection accident prediction and prevention possible. The intersection accident scenarios deduced for the purpose of the development of I-ADAS and self-driving system were analyzed regarding the severity of injuries and other factors. Based on these factors, possible methods to predict and prevent intersection accidents by using video footages and other data from the black box installed on cars were simply suggested.

Reflection removal on the windshield during nighttime driving

The in-vehicle black box camera has become a popular device in many countries for security monitoring and event capture. However, the content of the video is often degraded by the reflection appear on the windshield. The reflection is produced by light and dust outside. We focus on this problem and attempt to improve the readability of video content by removing the undesired reflection layer from the nighttime onboard video. We firstly apply gradient prior and the characteristics of reflection to automatically detect the reflection regions; the reflections in these regions are then classified into the major and minor reflection. Next, we suppress the major reflections by reducing the gradient strength of the reflection layer. Based on the different distribution of gradient histograms between the background layer and the reflection layer, we construct the objective function to remove the major reflection residue and minor reflection. In the reflection removal process, we introduce luminance and chrominance fidelity terms into the objective function to maintain the brightness and chrominance of the background layer. Finally, we obtain a reflection removal result. Evaluation of removal experiments on nighttime onboard images shows that our proposed algorithm may successfully remove the reflection layer on the windshield in the nighttime onboard video.

Efficient Vanishing Point Detection for Driving Assistance Based on Visual Saliency Map and Image Segmentation from a Vehicle Black-Box Camera

Techniques for detecting a vanishing point (VP) which estimates the direction of a vehicle by analyzing its relationship with surrounding objects have gained considerable attention recently. VPs can be used to support safe vehicle driving in areas such as for autonomous driving, lane-departure avoidance, distance estimation, and road-area detection, by detecting points in which parallel extension lines of objects are concentrated at a single point in a 3D space. In this paper, we proposed a method of detecting the VP in real time for applications to intelligent safe-driving support systems. In order to support safe driving of autonomous vehicles, it is necessary to drive the vehicle with the VP in center of the road image in order to prevent the vehicle from moving out of the road area while driving. Accordingly, in order to detect the VP in the road image, a method of detecting a point where straight lines intersect in an area where edge directional feature information is concentrated is required. The visual attention model and image segmentation process are applied to quickly identify candidate VPs in the area where the edge directional feature-information is concentrated and the intensity contrast difference is large. In the proposed method, VPs are detected by analyzing the edges, visual-attention regions, linear components using the Hough transform, and image segmentation results in an input image. Our experimental results have shown that the proposed method could be applied to safe-driving support systems.

Efficient Vehicle Detection and Distance Estimation Based on Aggregated Channel Features and Inverse Perspective Mapping from a Single Camera

In this paper a method for detecting and estimating the distance of a vehicle driving in front using a single black-box camera installed in a vehicle was proposed. In order to apply the proposed method to autonomous vehicles, it was required to reduce the throughput and speed-up the processing. To do this, the proposed method decomposed the input image into multiple-resolution images for real-time processing and then extracted the aggregated channel features (ACFs). The idea was to extract only the most important features from images at different resolutions symmetrically. A method of detecting an object and a method of estimating a vehicle’s distance from a bird’s eye view through inverse perspective mapping (IPM) were applied. In the proposed method, ACFs were used to generate the AdaBoost-based vehicle detector. The ACFs were extracted from the LUV color, edge gradient, and orientation (histograms of oriented gradients) of the input image. Subsequently, by applying IPM and transforming a 2D input image into 3D by generating an image projected in three dimensions, the distance between the detected vehicle and the autonomous vehicle was detected. The proposed method was applied in a real-world road environment and showed accurate results for vehicle detection and distance estimation in real-time processing. Thus, it was showed that our method is applicable to autonomous vehicles.

Patch-Based Crack Detection in Black Box Images Using Convolutional Neural Networks

Cracks cause deterioration of road performance and functional or structural failure if not managed in a timely manner. This paper proposes an automated crack detection method using a car black box camera to address this problem. The proposed method uses a deep learning model [i.e., convolutional neural network (CNN)] composed of segmentation and classification modules. The segmentation process is performed to extract only the road surface in order to remove elements that interfere with crack detection in the black box image. Then, cracks are detected through analysis of patch units within the extracted road surface. The proposed CNN architecture classifies the elements of the road surface into three categories (i.e., crack, road marking, and intact area) with 90.45% accuracy. The results of the proposed CNN architecture are better than those of previous studies.

Reflection Removal Under Fast Forward Camera Motion.

The image quality of an in-vehicle black box camera is often degraded by the reflections of internal objects, dirt, and dust on the windshield. In this paper, we propose a novel algorithm that simultaneously removes the reflections and small dirt artifacts from in-vehicle black box videos under fast forward camera motion. The algorithm exploits the spatiotemporal coherence of the reflection and dirt, which remain stationary relative to the fast-moving background. Unlike previous algorithms, the algorithm first separates stationary reflection and then restores the background scene. To this end, we propose an average image prior, thereby imposing spatiotemporal coherence. The separation model is a two-layer model composed of stationary and background layers, where different gradient sparsity distributions are utilized in a region-based manner. Motion compensation in postprocessing is proposed to alleviate layer jitter due to vehicle vibrations. In evaluation experiments, the proposed algorithm successfully extracts the stationary layer from several real and synthetic black box videos.

도로의 파손 상태를 자동관리하기 위한 동영상 기반 실시간 포트홀 탐지 시스템

Potholes are caused by the presence of water in the underlying soil structure, which weakens the road pavement by expansion and contraction of water at freezing and thawing temperatures. Recently, automatic pothole detection systems have been studied, such as vibration-based methods and laser scanning methods. However, the vibration-based methods have low detection accuracy and limited detection area. Moreover, the costs for laser scanning-based methods are significantly high. Thus, in this paper, we propose a new pothole detection system using a commercial black-box camera. Normally, the computing power of a commercial black-box camera is limited. Thus, the pothole detection algorithm should be designed to work with the embedded computing environment of a black-box camera. The designed pothole detection algorithm has been tested by implementing in a black-box camera. The experimental results are analyzed with specific evaluation metrics, such as sensitivity and precision. Our studies confirm that the proposed pothole detection system can be utilized to gather pothole information in real-time.

Pothole Detection System Using a Black-box Camera.

Aging roads and poor road-maintenance systems result a large number of potholes, whose numbers increase over time. Potholes jeopardize road safety and transportation efficiency. Moreover, they are often a contributing factor to car accidents. To address the problems associated with potholes, the locations and size of potholes must be determined quickly. Sophisticated road-maintenance strategies can be developed using a pothole database, which requires a specific pothole-detection system that can collect pothole information at low cost and over a wide area. However, pothole repair has long relied on manual detection efforts. Recent automatic detection systems, such as those based on vibrations or laser scanning, are insufficient to detect potholes correctly and inexpensively owing to the unstable detection of vibration-based methods and high costs of laser scanning-based methods. Thus, in this paper, we introduce a new pothole-detection system using a commercial black-box camera. The proposed system detects potholes over a wide area and at low cost. We have developed a novel pothole-detection algorithm specifically designed to work with the embedded computing environments of black-box cameras. Experimental results are presented with our proposed system, showing that potholes can be detected accurately in real-time.

Graphical Selection of Optical Materials Using an Expanded Athermal Glass Map and Considering the Housing Material for an Athermal and Achromatic Design

This paper presents a new graphical method for selecting a pair of optical glass and housing materials to simultaneously achromatize and athermalize a multilens system composed of many elements. To take into account the lens spacing and housing, we quantify the lens power, chromatic power, and thermal power by weighting the ratio of the paraxial ray height at each lens to them. In addition, we introduce the equivalent single lens and the expanded athermal glass map including a housing material. Even though a lens system is composed of many elements, we can simply identify a pair of glass and housing materials that satisfies the athermal and achromatic conditions. Applying this method to design a black box camera lens equipped with a 1/4-inch image sensor having a pixel width of $2{\mu}m$ , the chromatic and thermal defocusings are reduced to less than the depth of focus, over the specified ranges in temperature and frequency.