Nonparametric semantic segmentation for 3D street scenes

In this paper we propose a method to generate a large scale and accurate dense 3D semantic map of street scenes. A dense 3D semantic model of the environment can significantly improve a number of robotic applications such as autonomous driving, navigation or localisation. Instead of using offline trained classifiers for semantic segmentation, our approach employs a data-driven, nonparametric method to parse scenes which easily scale to a large environment and generalise to different scenes. We use stereo image pairs collected from cameras mounted on a moving car to produce dense depth maps which are combined into a global 3D reconstruction using camera poses from stereo visual odometry. Simultaneously, 2D automatic semantic segmentation using a nonparametric scene parsing method is fused into the 3D model. Furthermore, the resultant 3D semantic model is improved with the consideration of moving objects in the scene. We demonstrate our method on the publicly available KITTI dataset and evaluate the performance against manually generated ground truth.

Vision-only autonomous navigation using topometric maps

This paper presents a mapping and navigation system for a mobile robot, which uses vision as its sole sensor modality. The system enables the robot to navigate autonomously, plan paths and avoid obstacles using a vision based topometric map of its environment. The map consists of a globally-consistent pose-graph with a local 3D point cloud attached to each of its nodes. These point clouds are used for direction independent loop closure and to dynamically generate 2D metric maps for locally optimal path planning. Using this locally semi-continuous metric space, the robot performs shortest path planning instead of following the nodes of the graph --- as is done with most other vision-only navigation approaches. The system exploits the local accuracy of visual odometry in creating local metric maps, and uses pose graph SLAM, visual appearance-based place recognition and point clouds registration to create the topometric map. The ability of the framework to sustain vision-only navigation is validated experimentally, and the system is provided as open-source software.

Vision-based place recognition : how low can you go?

In this paper we use the algorithm SeqSLAM to address the question, how little and what quality of visual information is needed to localize along a familiar route? We conduct a comprehensive investigation of place recognition performance on seven datasets while varying image resolution (primarily 1 to 512 pixel images), pixel bit depth, field of view, motion blur, image compression and matching sequence length. Results confirm that place recognition using single images or short image sequences is poor, but improves to match or exceed current benchmarks as the matching sequence length increases. We then present place recognition results from two experiments where low-quality imagery is directly caused by sensor limitations; in one, place recognition is achieved along an unlit mountain road by using noisy, long-exposure blurred images, and in the other, two single pixel light sensors are used to localize in an indoor environment. We also show failure modes caused by pose variance and sequence aliasing, and discuss ways in which they may be overcome. By showing how place recognition along a route is feasible even with severely degraded image sequences, we hope to provoke a re-examination of how we develop and test future localization and mapping systems.

Vision, Viability and Value: Three perspectives on the performing arts across cultures, context and nations

How can we reach out to institutions, artists and audiences with sometimes radically different agendas to encourage them to see, participate in and support the development of new practices and programs in the performing arts? In this paper, based on a plenary panel at PSi#18 Performance Culture Industry at the University of Leeds, Clarissa Ruiz (Columbia), AnuradhaKapur (India) and Sheena Wrigley (England) together with interloctorBree Hadley (Australia) speak about their work in as policy-makers, managers and producers in the performing arts in Europe, Asia and America over the past several decades. Acknowledged trailblazers in their fields, Ruiz, Kapur and Wrigley all have a commitment to creating a vital, viable and sustainable performing arts ecologies. Each has extensive experience in performance, politics, and the challenging process of managing histories, visions, stakeholders, and sometimes scarce resources to generate lasting benefits for the various communities have worked for, with and within. Their work, cultivating new initiatives, programs or policy has made them expert at brokering relationships in and in between private, public and political spheres to elevate the status of and support for performing arts as a socially and economically beneficial activity everyone can participate in. Each gives examples from their own practice to provide insight into how to negotiate the interests of artistic, government, corporate, community and education partners, and the interests of audiences, to create aesthetic, cultural and / or economic value. Together, their views offer a compelling set of perspectives on the changing meanings of the ‘value of the arts’ and the effects this has had for the artists that make and arts organisations that produce and present work in a range of different regional, national and cross-national contexts.

Monocular vision based autonomous navigation for a cost-effective MAV in GPS-denied environments

In this paper, we present a monocular vision based autonomous navigation system for Micro Aerial Vehicles (MAVs) in GPS-denied environments. The major drawback of monocular systems is that the depth scale of the scene can not be determined without prior knowledge or other sensors. To address this problem, we minimize a cost function consisting of a drift-free altitude measurement and up-to-scale position estimate obtained using the visual sensor. We evaluate the scale estimator, state estimator and controller performance by comparing with ground truth data acquired using a motion capture system. All resources including source code, tutorial documentation and system models are available online.

Vision-based estimation of airborne target pseudobearing rate using hidden Markov model filters

The problem of estimating pseudobearing rate information of an airborne target based on measurements from a vision sensor is considered. Novel image speed and heading angle estimators are presented that exploit image morphology, hidden Markov model (HMM) filtering, and relative entropy rate (RER) concepts to allow pseudobearing rate information to be determined before (or whilst) the target track is being estimated from vision information.

Information and Learning Experiences in the Cube: Vision, Implementation and Interaction

Executive Summary This project has commenced an exploration of learning and information experiences in the QUT Cube. Understanding learning in this environment has the potential to inform current implementations and future project development. In this report, we present early findings from the first phase of an investigation into what makes learning possible in the context of a giant interactive multi-media display such as the QUT Cube, which is an award-winning configuration that hosts several projects.

Robust vision-based underwater homing using self-similar landmarks

Next-generation autonomous underwater vehicles (AUVs) will be required to robustly identify underwater targets for tasks such as inspection, localization, and docking. Given their often unstructured operating environments, vision offers enormous potential in underwater navigation over more traditional methods; however, reliable target segmentation often plagues these systems. This paper addresses robust vision-based target recognition by presenting a novel scale and rotationally invariant target design and recognition routine based on self-similar landmarks that enables robust target pose estimation with respect to a single camera. These algorithms are applied to an AUV with controllers developed for vision-based docking with the target. Experimental results show that the system performs exceptionally on limited processing power and demonstrates how the combined vision and controller system enables robust target identification and docking in a variety of operating conditions.

Characteristics of on-road driving performance of persons with central vision loss who use bioptic telescopes

Purpose. To compare the on-road driving performance of visually impaired drivers using bioptic telescopes with age-matched controls. Methods. Participants included 23 persons (mean age = 33 ± 12 years) with visual acuity of 20/63 to 20/200 who were legally licensed to drive through a state bioptic driving program, and 23 visually normal age-matched controls (mean age = 33 ± 12 years). On-road driving was assessed in an instrumented dual-brake vehicle along 14.6 miles of city, suburban, and controlled-access highways. Two backseat evaluators independently rated driving performance using a standardized scoring system. Vehicle control was assessed through vehicle instrumentation and video recordings used to evaluate head movements, lane-keeping, pedestrian detection, and frequency of bioptic telescope use. Results. Ninety-six percent (22/23) of bioptic drivers and 100% (23/23) of controls were rated as safe to drive by the evaluators. There were no group differences for pedestrian detection, or ratings for scanning, speed, gap judgments, braking, indicator use, or obeying signs/signals. Bioptic drivers received worse ratings than controls for lane position and steering steadiness and had lower rates of correct sign and traffic signal recognition. Bioptic drivers made significantly more right head movements, drove more often over the right-hand lane marking, and exhibited more sudden braking than controls. Conclusions. Drivers with central vision loss who are licensed to drive through a bioptic driving program can display proficient on-road driving skills. This raises questions regarding the validity of denying such drivers a license without the opportunity to train with a bioptic telescope and undergo on-road evaluation.

The effects and interactions of GABAergic and dopaminergic agents in the prevention of form deprivation myopia by brief periods of normal vision

Intravitreal injections of GABA antagonists, dopamine agonists and brief periods of normal vision have been shown separately to inhibit form-deprivation myopia (FDM). Our study had three aims: (i) establish whether GABAergic agents modify the myopia protective effect of normal vision, (ii) investigate the receptor sub-type specificity of any observed effect, and (iii) consider an interaction with the dopamine (DA) system. Prior to the period of normal vision GABAergic agents were applied either (i) individually, (ii) in combination with other GABAergic agents (an agonist with an antagonist), or (iii) in combination with DA agonists and antagonists. Water injections were given to groups not receiving drug treatments so that all experimental eyes received intravitreal injections. As shown previously, constant form-deprivation resulted in high myopia and when diffusers were removed for 2 h per day the period of normal vision greatly reduced the FDM that developed. GABA agonists inhibited the protective effect of normal vision whereas antagonists had the opposite effect. GABAA/C agonists and D2 DA antagonists when used in combination were additive in suppressing the protective effect of normal vision. A D2 DA agonist restored some of the protective effect of normal vision that was inhibited by a GABA agonist (muscimol). The protective effect of normal vision against form-deprivation is modifiable by both the GABAergic and DAergic pathways.

Vision in water

The purpose of this study is to determine visual performance in water, including the influence of pupil size. The water en-vironment was simulated by placing a goggle filled with saline in front of eyes, with apertures placed at the front of the goggle. Correction factors were determined for the different magnification under this condition in order to to estimate vision in water. Experiments were conducted on letter visual acuity (7 participants), grating resolution (8 participants), and grating contrast sensitivity (1 participant). For letter acuity, mean loss in vision in water, compared to corrected vision in air, varied between 1.1 log minutes of arc resolution (logMAR) for a 1mm aperture to 2.2 logMAR for a 7mm aperture. The vision in minutes of arc was described well by a linear relationship with pupil size. For grating acuity, mean loss varied between 1.1 logMAR for a 2mm aperture to 1.2 logMAR for a 6mm aperture. Contrast sensitivity for a 2mm aperture dete-riorated as spatial frequency increased, with 2 log unit loss by 3 cycles/degree. Superimposed on this deterioration were depressions (notches) in sensitivity, with the first three notches occurring at 0.45, 0.8 and 1.3 cycles/degree with esti-mates for water of 0.39, 0.70 and 1.13 cycles/degree. In conclusion, vision in water is poor. It becomes worse as pupil size increases, but the effects are much more marked for letter targets than for grating targets.

Vision in water

Purpose: To determine visual performance in water, including the influence of pupil size. Method: The water environment was simulated by placing a goggle filled with saline in front of eyes, with apertures placed at the front of the goggle. Correction factors were determined for the different magnification under this condition to estimate vision in water. Experiments were conducted on letter visual acuity (7 participants), grating resolution (8 participants), and grating contrast sensitivity (1 participant). Results: For letter acuity, mean loss in vision in water, compared to corrected vision in air, varied between 1.1 log minutes of arc resolution (logMAR) for a 1mm aperture to 2.2 logMAR for a 7mm aperture. The vision in minutes of arc was described well by a linear relationship with pupil size. For grating acuity, mean loss varied between 1.1 logMAR for a 2mm aperture to 1.2 logMAR for a 6mm aperture. Contrast sensitivity for a 2mm aperture deteriorated as spatial frequency increased, with 2 log unit loss by 3 cycles/degree. Superimposed on this deterioration were depressions (notches) in sensitivity, with the first three notches occurring at 0.45, 0.8 and 1.3 cycles/degree and with estimates for water of 0.39, 0.70 and 1.13 cycles/degree. Conclusion: Vision in water is poor. It becomes worse as pupil size increases, but the effects are much more marked for letter targets than for grating targets.

Translating vision research into policy and practice to improve the visibility, and hence safety, or road workers at night

The process of translating research into policy and practice is not well understood. This paper uses a case study approach to interpret an example of translation with respect to theoretical approaches identified in the literature. The case study concerns research into “biological motion” or “biomotion”: when lights are placed on the moveable joints of the body and the person moves in a dark setting, there is immediate and accurate recognition of the human form although only the lights can be seen. QUT was successful in gaining Australian Research Council funding with the support of the predecessors of the Queensland Department of Transport and Main Roads (TMR) to research the biomotion effect in road worker clothing using reflective tape rather than lights, and this resulted in the incorporation of biomotion marking into AS/NZS 4602.1 2011. The most promising approach to understanding the success of this translation, SWOV’s “knowledge utilisation approach” provided some insights but was more descriptive than predictive and provided “necessary but not sufficient” conditions for translation. In particular, the supportive efforts of TMR staff engaged in the review and promulgation of national standards were critical in this case. A model of the conclusions is presented. The experiences gained in this case should provide insights into the processes involved in effectively translating research into practice.

Crowd monitoring using computer vision

Novel computer vision techniques have been developed for automatic monitoring of crowed environments such as airports, railway stations and shopping malls. Using video feeds from multiple cameras, the techniques enable crowd counting, crowd flow monitoring, queue monitoring and abnormal event detection. The outcome of the research is useful for surveillance applications and for obtaining operational metrics to improve business efficiency.

HMM relative entropy rate concepts for vision-based aircraft manoeuvre detection

Machine vision is emerging as a viable sensing approach for mid-air collision avoidance (particularly for small to medium aircraft such as unmanned aerial vehicles). In this paper, using relative entropy rate concepts, we propose and investigate a new change detection approach that uses hidden Markov model filters to sequentially detect aircraft manoeuvres from morphologically processed image sequences. Experiments using simulated and airborne image sequences illustrate the performance of our proposed algorithm in comparison to other sequential change detection approaches applied to this application.