Inspection of pole-like structures using a vision-controlled VTOL UAV and shared autonomy

We present an approach for the inspection of vertical pole-like infrastructure using a vertical take-off and landing (VTOL) unmanned aerial vehicle and shared autonomy. Inspecting vertical structures, such as light and power distribution poles, is a time consuming, dangerous and expensive task with high operator workload. To address these issues, we propose a VTOL platform that can operate at close-quarters, whilst maintaining a safe stand-off distance and rejecting environmental disturbances. We adopt an Image based Visual Servoing (IBVS) technique using only two line features to stabilise the vehicle with respect to a pole. Visual, inertial and sonar data are used, making the approach suitable for indoor or GPS-denied environments. Results from simulation and outdoor flight experiments demonstrate the system is able to successfully inspect and circumnavigate a pole.

Short-data recursive HMM parameter estimation for rapid vision-based aircraft heading estimation

Rapid recursive estimation of hidden Markov Model (HMM) parameters is important in applications that place an emphasis on the early availability of reasonable estimates (e.g. for change detection) rather than the provision of longer-term asymptotic properties (such as convergence, convergence rate, and consistency). In the context of vision- based aircraft (image-plane) heading estimation, this paper suggests and evaluates the short-data estimation properties of 3 recursive HMM parameter estimation techniques (a recursive maximum likelihood estimator, an online EM HMM estimator, and a relative entropy based estimator). On both simulated and real data, our studies illustrate the feasibility of rapid recursive heading estimation, but also demonstrate the need for careful step-size design of HMM recursive estimation techniques when these techniques are intended for use in applications where short-data behaviour is paramount.

Assessing rod, cone and melanopsin contributions to human pupil flicker responses

Purpose: To determine the relative contributions of rods, cones and melanopsin to pupil responses in humans using temporal sinusoidal stimulation for light levels spanning the low mesopic to photopic range. Methods: A four-primary Ganzfeld photostimulator controlled flicker stimulations at seven light levels (-2.7 to 2 log cd/m2) and five frequencies (0.5 to 8Hz). Pupil diameter was measured using a high-resolution eyetracker. Three kinds of sinusoidal photoreceptor modulations were generated using silent substitution: 1) rod modulation, 2) cone modulation, and 3) combined rod and cone modulation in phase (Experiment 1) or phase shifted (Experiment 2) from a fixed rod phase. The melanopsin excitation was computed for each condition. A vector sum model was used to estimate the relative contribution of rods, cones and melanopsin to the pupil response. Results: From Experiment 1, the pupil frequency response peaked at 1Hz at two mesopic light levels for the three modulation conditions. Analyzing the rod-cone phase difference for the combined modulations (Experiment 2) identified a V-shaped response amplitude with a minimum between 135° and 180°. The pupil response phases increased as cone modulation phase increased. The pupil amplitude increased with increasing light level for cone and combined in-phase rod and cone modulation, but not for the rod modulation. Conclusions: These results demonstrate that cone- and rod-pathway contributions are more predominant than melanopsin contribution to the phasic pupil response. The combined rod, cone and melanopsin inputs to the phasic state of the pupil light reflex follow linear summation.

Towards Vision-Based Pose- and Condition-Invariant Place Recognition along Routes

Vision-based place recognition involves recognising familiar places despite changes in environmental conditions or camera viewpoint (pose). Existing training-free methods exhibit excellent invariance to either of these challenges, but not both simultaneously. In this paper, we present a technique for condition-invariant place recognition across large lateral platform pose variance for vehicles or robots travelling along routes. Our approach combines sideways facing cameras with a new multi-scale image comparison technique that generates synthetic views for input into the condition-invariant Sequence Matching Across Route Traversals (SMART) algorithm. We evaluate the system’s performance on multi-lane roads in two different environments across day-night cycles. In the extreme case of day-night place recognition across the entire width of a four-lane-plus-median-strip highway, we demonstrate performance of up to 44% recall at 100% precision, where current state-of-the-art fails.

Kiwi Forego Vision in the Guidance of Their Nocturnal Activities

Background In vision, there is a trade-off between sensitivity and resolution, and any eye which maximises information gain at low light levels needs to be large. This imposes exacting constraints upon vision in nocturnal flying birds. Eyes are essentially heavy, fluid-filled chambers, and in flying birds their increased size is countered by selection for both reduced body mass and the distribution of mass towards the body core. Freed from these mass constraints, it would be predicted that in flightless birds nocturnality should favour the evolution of large eyes and reliance upon visual cues for the guidance of activity. Methodology/Principal Findings We show that in Kiwi (Apterygidae), flightlessness and nocturnality have, in fact, resulted in the opposite outcome. Kiwi show minimal reliance upon vision indicated by eye structure, visual field topography, and brain structures, and increased reliance upon tactile and olfactory information. Conclusions/Significance This lack of reliance upon vision and increased reliance upon tactile and olfactory information in Kiwi is markedly similar to the situation in nocturnal mammals that exploit the forest floor. That Kiwi and mammals evolved to exploit these habitats quite independently provides evidence for convergent evolution in their sensory capacities that are tuned to a common set of perceptual challenges found in forest floor habitats at night and which cannot be met by the vertebrate visual system. We propose that the Kiwi visual system has undergone adaptive regressive evolution driven by the trade-off between the relatively low rate of gain of visual information that is possible at low light levels, and the metabolic costs of extracting that information.

Large-scale habitat mapping using vision-based AUVs : experiences, challenges & vehicle design

This paper overviews the development of a vision-based AUV along with a set of complementary operational strategies to allow reliable autonomous data collection in relatively shallow water and coral reef environments. The development of the AUV, called Starbug, encountered many challenges in terms of vehicle design, navigation and control. Some of these challenges are discussed with focus on operational strategies for estimating and reducing the total navigation error when using lower-resolution sensing modalities. Results are presented from recent field trials which illustrate the ability of the vehicle and associated operational strategies to enable rapid collection of visual data sets suitable for marine research applications.

Towards coordinated vision-based docking using an Autonomous Surface Vehicle

This paper describes the development and experimental evaluation of a novel vision-based Autonomous Surface Vehicle with the purpose of performing coordinated docking manoeuvres with a target, such as an Autonomous Underwater Vehicle, on the water’s surface. The system architecture integrates two small processor units; the first performs vehicle control and implements a virtual force obstacle avoidance and docking strategy, with the second performing vision-based target segmentation and tracking. Furthermore, the architecture utilises wireless sensor network technology allowing the vehicle to be observed by, and even integrated within an ad-hoc sensor network. The system performance is demonstrated through real-world experiments.

Vision-based operations of a large industrial vehicle : autonomous hot metal carrier

Hot metal carriers (HMCs) are large forklift-type vehicles used to move molten metal in aluminum smelters. This paper reports on field experiments that demonstrate that HMCs can operate autonomously and in particular can use vision as a primary sensor to locate the load of aluminum. We present our complete system but focus on the vision system elements and also detail experiments demonstrating reliable operation of the materials handling task. Two key experiments are described, lasting 2 and 5 h, in which the HMC traveled 15 km in total and handled the load 80 times.

Improving robot vision models for object detection through interaction

We propose a method for learning specific object representations that can be applied (and reused) in visual detection and identification tasks. A machine learning technique called Cartesian Genetic Programming (CGP) is used to create these models based on a series of images. Our research investigates how manipulation actions might allow for the development of better visual models and therefore better robot vision. This paper describes how visual object representations can be learned and improved by performing object manipulation actions, such as, poke, push and pick-up with a humanoid robot. The improvement can be measured and allows for the robot to select and perform the `right' action, i.e. the action with the best possible improvement of the detector.

Vision-based control for mining automation

The mining industry is highly suitable for the application of robotics and automation technology since the work is both arduous and dangerous. Visual servoing is a means of integrating noncontact visual sensing with machine control to augment or replace operator based control. This article describes two of our current mining automation projects in order to demonstrate some, perhaps unusual, applications of visual servoing, and also to illustrate some very real problems with robust computer vision

Special issue on robot vision

The International Journal of Robotics Research (IJRR) has a long history of publishing the state-of-the-art in the field of robotic vision. This is the fourth special issue devoted to the topic. Previous special issues were published in 2012 (Volume 31, No. 4), 2010 (Volume 29, Nos 2–3) and 2007 (Volume 26, No. 7, jointly with the International Journal of Computer Vision). In a closely related field was the special issue on Visual Servoing published in IJRR, 2003 (Volume 22, Nos 10–11). These issues nicely summarize the highlights and progress of the past 12 years of research devoted to the use of visual perception for robotics.

A roadmap for interpreting the literature on vision and driving

Over the past several decades there has been a sharp increase in the number of studies focused on the relationship between vision and driving. The intensified attention to this topic has most likely been stimulated by the lack of an evidence basis for determining vision standards for driving licensure and a poor understanding about how vision impairment impacts driver safety and performance. Clinicians depend on the literature on vision and driving to advise visually impaired patients appropriately about driving fitness. Policy makers also depend on the scientific literature in order to develop guidelines that are evidence-based and are thus fair to persons who are visually impaired. Thus it is important for clinicians and policy makers alike to understand how various study designs and measurement methods should be interpreted so that the conclusions and recommendations they make are not overly broad, too narrowly constrained, or even misguided. We offer a methodological framework to guide interpretations of studies on vision and driving that can also serve as a heuristic for researchers in the area. Here, we discuss research designs and general measurement methods for the study of vision as they relate to driver safety, driver performance, and driver-centered (self-reported) outcomes.

Fear of falling in vision impairment

Falls are the leading cause of injury-related morbidity and mortality among older adults. In addition to the resulting physical injury and potential disability after a fall, there are also important psychological consequences, including depression, anxiety, activity restriction, and fear of falling. Fear of falling affects 20 to 43% of community-dwelling older adults and is not limited to those who have previously experienced a fall. About half of older adults who experience fear of falling subsequently restrict their physical and everyday activities, which can lead to functional decline, depression, increased falls risk, and reduced quality of life. Although there is clear evidence that older adults with visual impairment have higher falls risk, only a limited number of studies have investigated fear of falling in older adults with visual impairment and the findings have been mixed. Recent studies suggest increased levels of fear of falling among older adults with various eye conditions, including glaucoma and age-related macular degeneration, whereas other studies have failed to find differences. Interventions, which are still in their infancy in the general population, are also largely unexplored in those with visual impairment. The major aims of this review were to provide an overview of the literature on fear of falling, its measurement, and risk factors among older populations, with specific focus on older adults with visual impairment, and to identify directions for future research in this area.