Capturing Motion
Conventional camera technologies have their limitations when trying to reconstruct the motion of an animal in flight. Cameras take in huge amounts of information that must then be converted into streams of coordinates later downstream, with much of the information that is captured thrown away. This often takes a large amount of human and machine time, to pick out the points of the animal that define its anatomy and to track them through images, multiplied by the number of cameras. Then there is the problem of gathering enough pixels to resolve any key features, keeping the animal in focus, or making sure there is enough light from the right direction.
A common darter dragonfly hovers whilst patrolling its territory. This kind of behaviour is hard to capture with conventional camera tracking.
Advances in image processing are already making inroads to this challenge, but for the present time, motion-capture systems offer an alternative to conventional camera systems that circumvent many of the latter's problems. It gives the coordinates of an objects position near instantly, with unparalleled accuracy, cutting out the time and energy spent digitising conventional footage. This means that much larger quantities of data can be gathered and stored allowing for more comprehensive analysis.
For these reasons, motion capture is used extensively in any industry that needs to record movement for analysis or replication, including film studios, game studios, and sports sciences. My current research aims are to improve the minituarisation of motion capture to insect spatial scales, allowing us to analyse insect flight like never before.
For these reasons, motion capture is used extensively in any industry that needs to record movement for analysis or replication, including film studios, game studios, and sports sciences. My current research aims are to improve the minituarisation of motion capture to insect spatial scales, allowing us to analyse insect flight like never before.
A common darter dragonfly with motion capture markers to track its head and body.
The Equipment
|
Motion capture works by bouncing invisible, infra-red light of markers placed on the body that is moving. In our case the body of an insect. Each camera has a ring of LEDs surrounding the lens that strobe infra-red light out into the scenery before it, relying on markers to bounce the light back to it. Multiple cameras operate at the same time from different angles to reconstruct marker positions in space.
|
|
|
The major limitation of motion capture are the markers. While we can ask a human and many vertebrates to wear markers either attached to clothing or harnesses, insects are often smaller and lighter than a single marker. Thus we have had to come up with new marker fabrication techniques to make the smallest motion capture markers in the world. These markers reflect light back along the path it came in on, meaning they are much brighter than their surroundings (from the perspective of our cameras).
|
A pair of black soldierflies wear 3x5 mm marker frames.
The Data
The accuracy of the data over a wide area is primarily what sets motion capture apart from competitor methodologies. We can locate the position of our special ento-markers to within 0.05mm. This accuracy allows us to reconstruct the exact orientation of parts of the animals anatomy, such as where the head is looking or the rotation of the body in space.
We can leave the motion capture system up and running for hours at a time due to the small data sizes, allowing us to get to the heart of what an insect is reacting too, above and beyond the analysis of any single path. We can also do this over metres cubed of volume, and can reconstruct the position of an object even when somewhat out of focus. All together this demonstrates that entomological motion capture is the best tool currently available to reconstruct otherwise elements of behaviour happening in mid air, under dark conditions, or faster than the human eye can percieve.
So what next?
Motion capture at the entomological scale seems the perfect means by which to answer one of the field's oldest questions. Why are moths (and other insects flying in low light levels) drawn to bright, artificial light-sources? Most studies on this subject have relied on indirect measurements or significant manipulation of the animal in order to track it's motion around a light-source in the dark. With infra-red motion capture, we can measure the movements of animals in the complete absence of visible light, and track the exacting motion with accuracy and certainty. Using this data we can fit various models to the flightpaths that the animals take, given us a more reliable impression of exactly what is happening to insects in the presence of artificial light.
To understand the problem, I'll discuss some of the models that could potentially be tested and what we might expect to see if they represent what is going on.
Direct Attraction
The simplest possible model, in which insects take a direct flightpath to the lightsource.
Direct attraction is supported by the idea that light to a flying animal represents open space. Just think of being in a forest and looking toward it's edge, the brightest areas are likely to be your clearest routes out. Pigeons have been shown to navigate through effective forests by steering for perceived brighter areas, so why not insects? This effect could be heightened if the bright light stimulates an escape response in the animal, increasing it's motivation to fly towards what is perceived to be the clearest escape path.
This explanation suffers somewhat when observing the behaviour of any insects clustering around a light-source. While insects are capable of flying in relatively direct paths between desired locations, they seldom appear to be doing so around an artificial light source. Thus, it's worth considering less-simple mechanisms that could be driving the response.
This explanation suffers somewhat when observing the behaviour of any insects clustering around a light-source. While insects are capable of flying in relatively direct paths between desired locations, they seldom appear to be doing so around an artificial light source. Thus, it's worth considering less-simple mechanisms that could be driving the response.
Indirect Attraction
|
As illustrated to the right, insects often appear to spiral when approaching a lightsource, taking a repetitive and indirect route as they approach. This is particularly the case when observing insects clustering under a street-light, for instance. Indirect attraction could be for many reasons, but one mechanism might work for multiple explanations. If the insect tries to keep the light source at a fixed point (in the front half of it's visual field), then it will gradually proceed in on an approximate logarithmic spiral toward the light-source. If the light-source were much, much further away (i.e. the moon as below), then keeping it at a fixed angle could help moths remain flight-stable and guide them in a straight line. However it's the proximity of the light (a feature extremely recent in evolutionary time) that causes the spiraling path. This is effectively deviated pursuit, for more on which, see my target interception page.
|
|
If insects tried to keep a light-source like the moon at a fixed point of their vision, it could help guide them in a straight line. If they tried the same thing with a local artificial light-source, it would drive them on a spiral path that approaches the light.
Flight Inversion
Flying is a complex art. An airborne animal needs to be able to detect which way is up, a challenge as unlike us here on the floor, they can't rely on the stabilising reaction-force of the ground. Insects detect their orientation and any changes to it in a variety of ways, but principal among them is using vision. The sky is generally brighter than the ground (under natural conditions), by at least several orders of magnitude. This is true even at night. However, this may not be the case when an artificial light-source is present. When the sun is up, any artificial light-source is unlikely to compare with how incredibly bright the sky is above, but at night this could well be subverted, at least within the local area near a light. Insects in flight will rotate themselves in the air so that the brightest region of the scenery is roughly on their backs, this is called the dorsal light response.
Any insect that is flying over a well-lit area at night may become confused, and think the ground is the sky. This may cause them to roll over. It doesn't take an experienced pilot to realise that turning your aircraft upside down might create problems. The lift forces the animal would normally use to counteract gravity will now add to it, causing the insect to crash. This effect would not need to work in isolation, but would result in night-flying insects being severely hampered in artificially lit areas at night.
If we determine that stability is an issue, there are further test required to help us understand the mechanism of the problem. One of which is which visual system is responsible for the animals attraction. Many flying insects operate two visual systems at the same time. The compound eyes of insects are well known and are there to resolve images and details. Lesser known is the ocellar sytem featuring up to 3 individual light-sensors call ocelli. These ocelli are camera-type eyes (like ours they use a single lens for a whole retina rather than many lenses as in compound eyes), but they are wildly out-of-focus. This is because they aren't generally there to detect images, but instead to operate as sensitive and fast light-sensors, accurately determining the bright from dark regions of the scenery around the animal. This high-speed makes them useful for maintaining the insect's balance, and thus they might play a role in the problems surrounding an artificial light. We can test this by covering either the ocelli or the compound eyes or both of an insect, and then seeing how each manipulation alters the insect's flight path and it's orientation towards a light.
Potential manipulations to an insect to determine how each of the two different visual systems (ocellar and compound) contribute to light-attraction behaviour.
Why this question is an important one
Insect populations are crashing all over the world. This is happening for a vast array of reasons, revolving around our human manipulation of the environment. These may include land-use changes, pesticides, climate change, and many other depressing factors. An oft ignored, but important, factor is also likely due to our pollution of the night skies. The amount of artificial light produced at night is one of the most salient features of any human dwelling, and we can often see cities from a long way off based on their hazy glow. This light is, on an evolutionary time-scale, and extremely recent phenomenon. Thus it's unsurprising that certain animals, such as night-flying insects may be struggling to adapt. While we might observe the clusters of insects wheeling under a lamppost, how many more have been driven off-course or had their behaviour-pattern thrown into disarray? By looking at the problem of insect light-attraction we might hope to get a better gauge, and in doing so may be able to better limit the effect that our night-lighting has on the environment.
