Obstacle detection using optical flow

Motion has been examined in biology to be a critical component for obstacle avoidance and navigation. In particular, optical flow is a powerful motion cue that has been exploited in many biological systems for survival. In this paper, we investigate an obstacle detection system that uses optical flow to obtain range information to objects. Our experimental results demonstrate that optical flow is capable of providing good obstacle information but has obvious failure modes. We acknowledge that our optical flow system has certain disadvantages and cannot be solely used for navigation. Instead, we believe that optical flow is a critical visual subsystem used when moving at reason- able speeds. When combined with other visual subsystems, considerable synergy can result.

Learning to avoid indoor obstacles from optical flow

Optical flow (OF) is a powerful motion cue that captures the fusion of two important properties for the task of obstacle avoidance − 3D self-motion and 3D environmental surroundings. The problem of extracting such information for obstacle avoidance is commonly addressed through quantitative techniques such as time-to-contact and divergence, which are highly sensitive to noise in the OF image. This paper presents a new strategy towards obstacle avoidance in an indoor setting, using the combination of quantitative and structural properties of the OF field, coupled with the flexibility and efficiency of a machine learning system.The resulting system is able to effectively control the robot in real-time, avoiding obstacles in familiar and unfamiliar indoor environments, under given motion constraints. Furthermore, through the examination of the networks internal weights, we show how OF properties are being used toward the detection of these indoor obstacles.

Deep Raman spectroscopy for the non-invasive standoff detection of concealed chemical threat agents

Deep Raman spectroscopy has been utilized for the standoff detection of concealed chemical threat agents from a distance of 15 meters under real life background illumination conditions. By using combined time and space resolved measurements, various explosive precursors hidden in opaque plastic containers were identified non-invasively. Our results confirm that combined time and space resolved Raman spectroscopy leads to higher selectivity towards the sub-layer over the surface layer as well as enhanced rejection of fluorescence from the container surface when compared to standoff spatially offset Raman spectroscopy. Raman spectra that have minimal interference from the packaging material and good signal-to-noise ratio were acquired within 5 seconds of measurement time. A new combined time and space resolved Raman spectrometer has been designed with nanosecond laser excitation and gated detection, making it of lower cost and complexity than picosecond-based laboratory systems.

The effect of topical adrenergic and anticholinergic agents on the choroidal thickness of young healthy adults

The human choroid is capable of rapidly changing its thickness in response to a variety of stimuli. However little is known about the role of the autonomic nervous system in the regulation of the thickness of the choroid. Therefore, we investigated the effect of topical parasympatholytic and sympathomimetic agents upon the choroidal thickness and ocular biometrics of young healthy adult subjects. Fourteen subjects (mean age 27.9 ± 4 years) participated in this randomized, single-masked, placebo-controlled study. Each subject had measurements of choroidal thickness (ChT) and ocular biometrics of their right eye taken before, and then 30 and 60 min following the administration of topical pharmacological agents. Three different drugs: 2% homatropine hydrobromide, 2.5% phenylephrine hydrochloride and a placebo (0.3% hydroxypropyl methylcellulose) were tested in all subjects; each on different days (at the same time of the day) in randomized order. Participants were masked to the pharmacological agent being used at each testing session. The instillation of 2% homatropine resulted in a small but significant increase in subfoveal ChT at 30 and 60 min after drug instillation (mean change 7 ± 3 μm and 14 ± 2 μm respectively; both p < 0.0001). The parafoveal choroid also exhibited a similar magnitude, significant increase in thickness with time after 2% homatropine (p < 0.001), with a mean change of 7 ± 0.3 μm and 13 ± 1 μm (in the region located 0.5 mm from the fovea center), 6 ± 1 μm and 12.5 ± 1 μm (1 mm from the fovea center) and 6 ± 2 μm and 12 ± 2 μm (1.5 mm from the fovea center) after 30 and 60 min respectively. Axial length decreased significantly 60 min after homatropine (p < 0.01). There were also significant changes in lens thickness (LT) and anterior chamber depth (ACD) (p < 0.05) associated with homatropine instillation. No significant changes in choroidal thickness, or ocular biometrics were found after 2.5% phenylephrine or placebo at any examination points (p > 0.05). In human subjects, significant increases in subfoveal and parafoveal choroidal thickness occurred after administration of 2% homatropine and this implies an involvement of the parasympathetic system in the control of choroidal thickness in humans.

SERS-barcoded colloidal gold NP assemblies as imaging agents for use in biodiagnostics

There is a growing need for new biodiagnostics that combine high throughput with enhanced spatial resolution and sensitivity. Gold nanoparticle (NP) assemblies with sub-10 nm particle spacing have the benefits of improving detection sensitivity via Surface enhanced Raman scattering (SERS) and being of potential use in biomedicine due to their colloidal stability. A promising and versatile approach to form solution-stable NP assemblies involves the use of multi-branched molecular linkers which allows tailoring of the assembly size, hot-spot density and interparticle distance. We have shown that linkers with multiple anchoring end-groups can be successfully employed as a linker to assemble gold NPs into dimers, linear NP chains and clustered NP assemblies. These NP assemblies with diameters of 30-120 nm are stable in solution and perform better as SERS substrates compared with single gold NPs, due to an increased hot-spot density. Thus, tailored gold NP assemblies are potential candidates for use as biomedical imaging agents. We observed that the hot-spot density and in-turn the SERS enhancement is a function of the linker polymer concentration and polymer architecture. New deep Raman techniques like Spatially Offset Raman Spectroscopy (SORS) have emerged that allow detection from beneath diffusely scattering opaque materials, including biological media such as animal tissue. We have been able to demonstrate that the gold NP assemblies could be detected from within both proteinaceous and high lipid containing animal tissue by employing a SORS technique with a backscattered geometry.