Older observers' tolerance of optical blur: age differences in the identification of defocused text signs

To determine if visual aging affects the ability to identify blurred text signs in daytime or nighttime viewing conditions, the photopic and mesopic legibility thresholds of young and old adult observers were compared at three levels of optically induced acuity (Experiment 1). For both age groups, legibility was reduced by nighttime luminance and degraded acuity. Surprisingly, older observers were better than younger ones in identifying defocused (optically blurred) text in both daytime and nighttime conditions. In Experiment 2, older observers were also superior to younger ones in identifying defocused standard and spatially matched novel text signs. These findings indicate that older observers' tolerance of optical blur is a generic ability, rather than one explained by familiarity with the low-pass optical profile of specific signs. Consistent with the notion that factors beyond acuity contribute more importantly to text legibility for older than for younger observers, acuity was a stronger predictor of legibility thresholds in the younger group. Actual or potential applications of this research include the need to consider functional acuity in visual screening protocols and the design of text displays, as well as the utilization of compensatory letter-recognition training for older observers or others with diminished acuity.     

Age-related difference in steering control under reduced visibility conditions

The current study investigated age-related differences in a steering control task under low visibility conditions. Younger and older drivers were presented with displays simulating forward vehicle motion through a 3D scene of random dots on a ground plane. The lateral position of the vehicle was perturbed by a simulated side wind gust according to a sum of sinusoidal functions. The drivers’ task was to steer the vehicle to maintain a straight path. The visibility of the driving scene was reduced by reducing the quantity and the quality of the optical flow field. We found that performance decreased when visibility was reduced for both older and younger drivers, with better performance for younger drivers as compared with older drivers. An age-related interaction was also found with deteriorated optical flow information. These results suggest that under reduced visibility conditions, older drivers may have increased accident risks due to decreased ability to successfully steer the vehicle.     

面向智能驾驶的平行视觉感知：基本概念、框架与应用

目的 视觉感知技术是智能车系统中的一项关键技术,但是在复杂挑战下如何有效提高视觉性能已经成为智能驾驶领域的重要研究内容。本文将人工社会（artificial societies）、计算实验（computational experiments）和平行执行（parallel execution）构成的ACP方法引入智能驾驶的视觉感知领域,提出了面向智能驾驶的平行视觉感知,解决了视觉模型合理训练和评估问题,有助于智能驾驶进一步走向实际应用。方法 平行视觉感知通过人工子系统组合来模拟实际驾驶场景,构建人工驾驶场景使之成为智能车视觉感知的“计算实验室”;借助计算实验两种操作模式完成视觉模型训练与评估;最后采用平行执行动态优化视觉模型,保障智能驾驶对复杂挑战的感知与理解长期有效。结果 实验表明,目标检测的训练阶段虚实混合数据最高精度可达60.9%,比单纯用KPC（包括：KITTI（Karlsruhe Institute of Technology and Toyota Technological Institute）,PASCAL VOC（pattern analysis,statistical modelling and computational learning visual object classes）和MS COCO（Microsoft common objects in context））数据和虚拟数据分别高出17.9%和5.3%;在评估阶段相较于基准数据,常规任务（-30°且垂直移动）平均精度下降11.3%,环境任务（雾天）平均精度下降21.0%,困难任务（所有挑战）平均精度下降33.7%。结论 本文为智能驾驶设计和实施了在实际驾驶场景难以甚至无法进行的视觉计算实验,对复杂视觉挑战进行分析和评估,具备加强智能车在行驶过程中感知和理解周围场景的意义。     

Laser projection versus a CRT display in the visual perception of aircraft aspect

High-resolution visual displays have been designed for flight simulation so that observers may judge the aspect angle of aircraft at far distances. The present experiment compares two display devices as untrained observers judge the spatial orientation of two target aircraft: F-15 and F-16 jets. The display devices are a prototype direct-write microlaser projector and an SXGA-format CRT display. Observers' accuracy of aircraft identification is better with the laser projector, and recognition response times are faster. A simple rule was found to fit the observers' response times; it is expressed in terms of a statistic on the autocorrelation of black-and-white silhouette images of aircraft. Observers' estimates of aspect are biased by the laser projector, whereas observers' estimates of aspect are accurate on average with the SXGA display. This bias in estimation of aspect may be attributable to variations in line brightness introduced by the laser projector. Actual or potential applications of this research include the evaluation of high-resolution visual displays for the training of basic fighter maneuvers with military jet pilots.     

Positive display polarity is advantageous for both younger and older adults

The effect of display polarity on visual acuity and proofreading performance was investigated for younger and older adults. An advantage of positive polarity (dark characters on light background) over negative polarity (light characters on dark background) was expected for younger adults, but the effects on older adults were ambiguous. Light scatter due to residues in the senescent lens and vitreous humour could reverse the typical advantage of positive polarity. However, age-related changes lead to a decline in retinal illuminance. Brighter positive polarity displays should help to compensate for this decline and, accordingly, lead to better performance than darker negative polarity displays. Participants conducted a visual acuity test with black optotypes on white background or white optotypes on black background and performed a proofreading task in the same polarity. A positive polarity advantage was found for both age groups. The presentation in positive polarity is recommended for all ages.

Practitioner summary: In an ageing society, age-related vision changes need to be considered when designing digital displays. Visual acuity testing and a proofreading task revealed a positive polarity advantage for younger and older adults. Dark characters on light background lead to better legibility and are strongly recommended independent of observer's age.     

The art of descrying distance

How naive can one experiment be? Imagine asking observers to judge distance -- yes, literally asking them. And in asking them, no visual evoked potentials were measured, and changes in blood flow to the occipital cortex were ignored. But serendipity occurs even to the well-prepared: One's best thought for a new experiment can prove to be one's first unbiased, uncomplicated thought.Here's a cursory review of the events: A number of observers were asked to judge distances between marked posts in a field at night. They made their judgments with the help of night vision goggles (NVG). At first the observers underestimated the distances on average. Then an experimenter began to correct the observers after each judgment. The observers' judgments of distance became accurate. On average, the observers began to estimate those distances accurately. Is this and incredible story or a difficult one?     

Visual perception modelling for intelligent virtual driver agents in synthetic driving simulation

This paper introduces a vision model used within a synthetic (or purely software-based) driving simulation framework. This framework represents driver decision-making, individual vehicle movement and emergent traffic flow, and is intended to aid the integration of driver psychology, traffic management and vehicle engineering. The aims of developing the vision model discussed here are twofold: Firstly, to remove the unrealistic availability of ‘perfect knowledge’ concerning the positions and velocities of vehicles in a simulation and secondly, to provide a means of introducing deeper cognitive models of driver reasoning and behaviour. The paper presents the essential mechanisms of the vision model, along with the results of initial validation experiments conducted with Leicestershire Constabulary, Traffic Division in the UK. In these experiments, subjects' visual perception of positions and speeds of moving vehicles were measured and compared with estimations of the agent based driving simulator. The results have demonstrated the feasibility of modelling driver vision within an agent based traffic simulation using concepts derived from AI and ALife systems. The paper is completed by a short discussion on the future development of cognitive models enabled through more detailed vision modelling.     

A VE framework to study visual perception and action

In the real world, vision operates in harmony with self-motion yielding the observer to unambiguous perception of the three-dimensional (3D) space. In laboratory conditions, because of technical difficulties, researchers studying 3D perception have often preferred to use the substitute of a stationary observer, somehow neglecting aspects of the action-perception cycle. Recent results in visual psychophysics have proved that self-motion and visual processes interact, leading the moving observer to interpret a 3D virtual scene differently from a stationary observer. In this paper we describe a virtual environment (VE) framework which presents very interesting characteristics for designing experiments in visual perception during action. These characteristics arise in a number of ways from the design of a unique motion capture device. First, its accuracy and the minimal latency in position measurement; second, its ease of use and the adaptability to different display interfaces. Such a VE framework enables the experimenter to recreate stimulation conditions characterised by a degree of sensory coherence typical of the real world. Moreover, because of its accuracy and flexibility, the same device can be used as a measurement tool to perform elementary but essential calibration procedures. The VE framework has been used to conduct two studies which compare the perception of 3D variables of the environment in moving and in stationary observers under monocular vision. The first study concerns the perception of absolute distance, i.e. the distance separating an object and the observer. The second study refers to the perception of the orientation of a surface both in the absence and presence of conflicts between static and dynamic visual cues. In the two cases, the VE framework has enabled the design of optimal experimental conditions, permitting light to be shed on the role of action in 3D visual perception.     

Effects of age and illumination on night driving: a road test

OBJECTIVE: This study investigated the effects of drivers' age and low light on speed, lane keeping, and visual recognition of typical roadway stimuli. BACKGROUND: Poor visibility, which is exacerbated by age-related changes in vision, is a leading contributor to fatal nighttime crashes. There is little evidence, however, concerning the extent to which drivers recognize and compensate for their visual limitations at night. METHOD: Young, middle-aged, and elder participants drove on a closed road course in day and night conditions at a "comfortable" speed without speedometer information. During night tests, headlight intensity was varied over a range of 1.5 log units using neutral density filters. RESULTS: Average speed and recognition of road signs decreased significantly as functions of increased age and reduced illumination. Recognition of pedestrians at night was significantly enhanced by retroreflective markings of limb joints as compared with markings of the torso, and this benefit was greater for middle-aged and elder drivers. Lane keeping showed nonlinear effects of lighting, which interacted with task conditions and drivers' lateral bias, indicating that older drivers drove more cautiously in low light. CONCLUSION: Consistent with the hypothesis that drivers misjudge their visual abilities at night, participants of all age groups failed to compensate fully for diminished visual recognition abilities in low light, although older drivers behaved more cautiously than the younger groups. APPLICATION: These findings highlight the importance of educating all road users about the limitations of night vision and provide new evidence that retroreflective markings of the limbs can be of great benefit to pedestrians' safety at night.     

视觉感知的端到端自动驾驶运动规划综述

视觉感知模块能够利用摄像机等视觉传感器获取丰富的图像和视频信息,进而检测自动驾驶汽车视野中的车辆、行人与交通标识等信息,是自动驾驶最有效、成本最低的感知方式之一。运动规划为自主车辆提供从车辆初始状态到目标状态的一系列运动参数和驾驶动作,而端到端的模型能够直接从感知的数据获取车辆的运动参数,因而受到广泛的关注。为了全面反映视觉感知的端到端自动驾驶运动规划方法的研究进展,本文对国内外公开发表的具有代表性和前沿的论文进行了概述。首先分析端到端方法的应用,以及视觉感知和运动规划在端到端自动驾驶中的作用,然后以自主车辆的学习方式作为分类依据,将视觉感知的端到端自动驾驶运动规划的实现方法分为模仿学习和强化学习两大类,并对各类方法的不同算法进行了归纳和分析;考虑到现阶段端到端模型的研究面临着虚拟到现实的任务,故对基于迁移学习的方法进行了梳理。最后列举与自动驾驶相关的数据集和仿真平台,总结存在的问题和挑战,对未来的发展趋势进行思考和展望。视觉感知的端到端自动驾驶运动规划模型的普适性强且结构简单,这类方法具有广阔的应用前景和研究价值,但是存在不可解释和难以保证绝对安全的问题,未来需要更多的研究改善端到端模型存在的局限性。     

Bayesian network model of overall print quality: Construction and structural optimisation

Prediction of overall visual quality based on instrumental measurements is a challenging task. Despite the several proposed models and methods, there exists a gap between the instrumental measurements of print and human visual assessment of natural images. In this work, a computational model for representing and quantifying the overall visual quality of prints is proposed. The computed overall quality should correspond to the human visual quality perception when viewing the printed images. The proposed model is a Bayesian network which connects the objective instrumental measurements to the subjective opinion distribution of human observers. This relationship can be used to score printed images, and additionally, to computationally study the connections of the attributes. A novel graphical learning approach using an iterative evolve-estimate-simulate loop learning the quality model based on psychometric data and instrumental measurements is suggested. The network structure is optimised by applying evolutionary computation (evolve). The estimation of the Bayesian network parameters is within the evolutionary loop. In this loop, the maximum likelihood approach is used (estimate). The stochastic learning process is guided by priors devised from the psychometric subjective experiments (performance through simulation). The model reveals and represents the explanatory factors between its elements providing insight to the psychophysical phenomenon of how observers perceive visual quality and which measurable entities affect the quality perception. By using true data, the design choices are demonstrated. It is also shown that the best-performing network establishes a clear and intuitively correct structure between the objective measurements and psychometric data.     

Distance estimation with night vision goggles: a little feedback goes a long way

Immediate feedback was given to correct observers' estimates of distance in an experiment in which those estimates were made outdoors at night while observers wore night vision goggles (NVGs). Initially observers made unguided estimates of distances between marked positions in an open field. Those distances ranged from 7.6 m (25 ft) to 64 m (210 ft). Later the same observers made more estimates. After each of these they were told the measured distance between the positions. During this training, the observers' height from the ground plane was either at a standing position or at an elevated position raised 2.3 m (7 ft 7 in) from standing position. After the training--either immediately after, a week later, or at both times--observers made unguided estimates of distance for a second time. These latter estimates of ground distance made with the NVGs were improved. Average improvement of the observers' estimates persisted for at least one week after training. This training can be applied to improve clearance estimates and estimates of hover height for pilots of rotary-wing aircraft.     

In‐water visual ship hull inspection using a hover‐capable underwater vehicle with stereo vision

Underwater visual inspection is an important task for checking the structural integrity and biofouling of the ship hull surface to improve the operational safety and efficiency of ships and floating vessels. This paper describes the development of an autonomous in‐water visual inspection system and its application to visual hull inspection of a full‐scale ship. The developed system includes a hardware vehicle platform and software algorithms for autonomous operation of the vehicle. The algorithms for vehicle autonomy consist of the guidance, navigation, and control algorithms for real‐time and onboard operation of the vehicle around the hull surface. The environmental perception of the developed system is mainly based on optical camera images, and various computer vision and optimization algorithms are used for vision‐based navigation and visual mapping. In particular, a stereo camera is installed on the underwater vehicle to estimate instantaneous surface normal vectors, which enables high‐precision navigation and robust visual mapping, not only on flat areas but also over moderately curved hull surface areas. The development process of the vehicle platform and the implemented algorithms are described. The results of the field experiment with a full‐scale ship in a real sea environment are presented to demonstrate the feasibility and practical performance of the developed system.     

基于视觉的三维重建关键技术研究综述

三维重建在视觉方面具有很高的研究价值, 在机器人视觉导航、智能车环境感知系统以及虚拟现实中被广泛应用.本文对近年来国内外基于视觉的三维重建方法的研究工作进行了总结和分析, 主要介绍了基于主动视觉下的激光扫描法、结构光法、阴影法以及TOF (Time of flight)技术、雷达技术、Kinect技术和被动视觉下的单目视觉、双目视觉、多目视觉以及其他被动视觉法的三维重建技术, 并比较和分析这些方法的优点和不足.最后对三维重建的未来发展作了几点展望.     

基于视觉模型的车型动态定位

车型识别系统(VehicleRecognitionSystem简称VRS)是智能交通的一个重要课题,文章根据实时提供的汽车图像进行了大量的研究,采用了基于视觉模型的自适应的汽车车体定位算法。该算法采用的是HSI彩色图像空间,HSI彩色空间更适合人的视觉。同时根据基于人眼分辨率的视觉模型自适应地进行汽车彩色图像的边缘检测,根据边缘的垂直投影图和水平投影图来定位车体的左右边界和上下边界,实验结果表明该算法具有较快的速度,完全能满足适时的车型识别系统的要求。     

IVVI 2.0: An intelligent vehicle based on computational perception

This paper presents the IVVI 2.0 a smart research platform to foster intelligent systems in vehicles. Computational perception in intelligent transportation systems applications has advantages, such as huge data from vehicle environment, among others, so computer vision systems and laser scanners are the main devices that accomplish this task. Both have been integrated in our intelligent vehicle to develop cutting-edge applications to cope with perception difficulties, data processing algorithms, expert knowledge, and decision-making. The long-term in-vehicle applications, that are presented in this paper, outperform the most significant and fundamental technical limitations, such as, robustness in the face of changing environmental conditions. Our intelligent vehicle operates outdoors with pedestrians and others vehicles, and outperforms illumination variation, i.e.: shadows, low lighting conditions, night vision, among others. So, our applications ensure the suitable robustness and safety in case of a large variety of lighting conditions and complex perception tasks. Some of these complex tasks are overcome by the improvement of other devices, such as, inertial measurement units or differential global positioning systems, or perception architectures that accomplish sensor fusion processes in an efficient and safe manner. Both extra devices and architectures enhance the accuracy of computational perception and outreach the properties of each device separately.     

Age and visual impairment decrease driving performance as measured on a closed-road circuit

In this study the effects of visual impairment and age on driving were investigated and related to visual function. Participants were 139 licensed drivers (young, middle-aged, and older participants with normal vision, and older participants with ocular disease). Driving performance was assessed during the daytime on a closed-road driving circuit. Visual performance was assessed using a vision testing battery. Age and visual impairment had a significant detrimental effect on recognition tasks (detection and recognition of signs and hazards), time to complete driving tasks (overall course time, reversing, and maneuvering), maneuvering ability, divided attention, and an overall driving performance index. All vision measures were significantly affected by group membership. A combination of motion sensitivity, useful field of view (UFOV), Pelli-Robson letter contrast sensitivity, and dynamic acuity could predict 50% of the variance in overall driving scores. These results indicate that older drivers with either normal vision or visual impairment had poorer driving performance compared with younger or middle-aged drivers with normal vision. The inclusion of tests such as motion sensitivity and the UFOV significantly improve the predictive power of vision tests for driving performance. Although such measures may not be practical for widespread screening, their application in selected cases should be considered.     

车路协同路侧感知融合方法的研究

随着智能网联汽车技术和产业的不断发展,智能网联汽车逐渐成为人们交通出行的选择之一。但受智能网联汽车自身环境感知系统对特定道路交通场景信息处理的局限,无法实现在所有行驶工况下安全高效的运行,其需车路协同路侧感知技术的辅助方能更安全高效的运行。海量的车路协同感知数据是城市道路和高速公路车路协同、运行分析和科学管理的宝藏,理解和分析这些数据是车路协同路侧感知融合的关键。面对车路协同路侧多传感器的不同数据,如何高效准确地挖掘和提取雷达、视频在不同时间、不同空间维度的数据,实现对重点交通场景(如视野盲区、急转弯道、隧道、桥梁)和交通事件、环境、设施安全等的雷达、视频数据进行快速融合检测、识别与检索,通过蜂窝车联网C-V2X网络在一定时延范围内有效地将路侧感知融合结果数据发送给智能网联汽车,确保其安全高效的行驶,是面向智能网联汽车辅助驾驶的车路协同路侧感知融合的关键问题。基于智能网联汽车其自身环境感知能力,对道路智能基础设施感知网络中的多传感器融合方法进行研究分析,提出了基于误差方差的多传感器融合算法,与非智能道路相比,其效率更高,更加智能化,可有效解决道路交通运行环境中存在的常见问题,为人们提供更加安全、高效、优质的交通出行服务。     

Ordinal depth information from accommodation?

The ability to judge egocentric distance was assessed in two groups of six observers using a manual pointing task. The purpose of the study was to determine the extent to which blur-driven accommodation can provide information on target distance in the absence of any retinal cues to distance. Observers were extremely accurate when carrying out the pointing task in a 'full-cue' condition. In contrast, observers were extremely poor at carrying out the task when accommodation was the only distance cue available. Responses on individual trials bore little relationship to the actual target distance in any of the observers. On the other hand, accommodation weakly biased the mean responses in some observers. This bias appears to be due to the observers' effective use of accommodation to determine whether the target presented in one trial was nearer or further away than the target presented in the previous trial. Accommodation therefore appears to provide ordinal information, although the distance signal may actually arise from accommodation-driven vergence. The poverty of accommodation as a source of metric information was highlighted in a second group of observers who all demonstrated a strong bias when perceiving distance in the presence of an initially ambiguous retinal cue. It is concluded that accommodation can act as a source of ordinal distance information in the absence of other cues to distance but the contribution of accommodation to normal distance perception in full-cue conditions is questioned.     

Comparative study of the effects of auditory, visual and multimodality displays on drivers' performance in advanced traveller information systems

A simulator study was conducted to compare 16 younger (mean age 22 years) and 16 older (mean age 68 years) drivers' ratings of workload (time, visual, psychological stress) and performance of navigation and button-pushing (identification of vehicle or road hazards) tasks under both high- and low-load driving conditions when simple or complex advanced traveller information (ATI) was presented visually only, aurally only or by multimodality (visual and auditory) display. For all participants, both the auditory and multimodality displays produced better performance in terms of response times, total number of correct turns and subjective workload ratings than those of using the visual-only display. Participants using the multimodality display also made the fewest errors related to push-button and navigation tasks, and controlled their vehicles properly. The visual display led to less safe driving, apparently because it imposed higher demands on the drivers' attention. An age effect was found in the present study, with younger drivers performing better and reporting less stress than older drivers. Notably, however, use of the multimodality display significantly improved the older drivers' performance in the button-pushing task.