Geospatial Technology Used to Image the Human Eye
BroadSpot of San Francisco is developing a miniaturized camera to look at the inside of the retina. It turns out, the compact form factor technology for imaging the inside of a concave surface like an eye has a lot in common with what is used in the geospatial industry.
Dr. Tushar Ranchod, president and CEO of BroadSpot, says the Arched Retinal Camera Array (ARCA) technology created by Visual Intelligence is a good tool for imaging the human eye. It is a wide-field camera technology with a design that can accommodate low cost sensors.
“Typically, it is hard to image a wide field through a very small opening the size of a pupil. In an eye, using a single wide field imaging pathway creates lots of optical distortion at the edges of the image since it is a concave surface. Using an array of sensors creates multiple overlapping images, each relatively free of distortion, so the composite image is high resolution and also relatively free of distortion. The fact that the sensors in the ARCA technology are pointed through a single small opening makes it a good choice for taking multiple images of a very wide area inside,” Ranchod says.
Imaging with Arrays
Visual Intelligence of Houston sells a line of airborne image collection systems used for large area collection, stereo, and oblique and 3D modeling. The technology BroadSpot licensed is currently used to image very large areas in airborne digital cameras for spatial mapping, and until recently the only industry using this technology.
Array cameras create high-resolution, high-quality images using array of imaging sensors to take multiple images simultaneously. A few other manufacturers have array cameras for consumers and cellphones using coplanar arrays, with sensors that are arranged next to each other with layouts like 4x4 or 10x10 and point in the same direction.
Pelican Imaging’s mobile plenoptic camera is one example, and Lytro’s a light-field camera is another. Lytro’s camera allows users to shift perspective using software to selectively focus different sections of the image by choosing an image from a single sensor or group of pixels after the photo is taken.
Visual Intelligence’s wide-field cameras are different because they use non-coplanar arrays. The numerous imaging sensors don’t point in the same direction, they point across each other, through a very small opening. They allow users to gather a lot of data in a single snap, and make it easier to digitally capture wide areas below through smaller holes in an aircraft in airborne mapping.
Two of the most important benefits of this approach are that very economical sensors can now be used, and a variety of types of sensors can be used simultaneously. With advanced synchronization controls, they behave like a single camera and generate a single image with a large field of view and good depth, geometry and radiometry, very economically.
Dr. Armando Guevara, president and CEO of Visual Intelligence, says, “One array can collect color imagery; one can collect thermal imagery; one can collect NIR—all looking through a very small hole. The array geometry allows for functional scalability and collection scalability. Users aren’t forced to buy a single sensor that does only one thing. They can scale and grow with it. Our mission for the digital camera industry is to create a platform that could be ubiquitous.”
“The virtue of this sensor family is that it’s all one sensor system. Typically in the mapping industry, a company buys one camera for oblique imaging, and when it needs a large format camera or 3D imaging, it has to buy another. Some of the larger surveying companies own at least 15 digital cameras from different manufacturers. That’s a major headache in costs, maintenance and learning to operate.”
Visual Intelligence is miniaturizing the array technology now to go into UAVs and is considering putting it into handheld devices, cellphones and cameras in the field.
Lower Cost Imaging
There are a few ways to image the retina, and BroadSpot is developing a family of products to do this, Ranchod explains. The most basic is color photography, and another uses special cameras, filters and Fluorescien dye to look at blood flow. “Our approach, a single camera with an array of imaging sensors, would be a major improvement for optometrists and ophthalmologists, offering a cost effective option to replace current imaging cameras.”
“Using an array approach is revolutionary,” Ranchod says. “It allows us to create a device that’s an order of magnitude less expensive than monolithic cameras, which are large and not portable. In the short term, this approach complements the existing technology. But the longer term trend is to eventually replace it.”
Portable and Economical: Keys to Creating Content-Driven Diagnostic Equipment
Guevara believes the $1M price of a typical camera stagnates the geospatial industry. Digital obsolescence also takes effect the moment a user buys one. It’s the same in the medical industry, and BroadSpot’s new camera is a good example of where both industries could go.
“Users can take this small device directly to patients, who no longer need to move around as they would with monolithic imaging devices. Making high quality devices more economical and coupling them with portability allows everyone to buy them,” Guevara says.
However, the power comes when combining this with content-driven diagnostics. After a scan is taken and cross referenced with a database with even more data about a patient, extra diagnostics can be made.
“If users organize information and create knowledge with expert and rule-based systems, they can bring in other information about a patient’s body,” Guevara says. “Maybe they can track where he lived to see if his eye problem is because he lives in an area that had airborne aerosols. Tying in a content approach creates 'the internet of everything.'”
BroadSpot is in the early development stage, and has obtained an exclusive license for this technology from Visual Intelligence. Ranchod says, “A number of ophthalmic experts have validated this approach and there is no question whether the concept can work. We are making iterative improvements to manufactured devices to refine the technology and provide proof-of-concept that this low cost concept makes sense and is the way to go.”
In the foreground, the retinal imaging device is touching the ocular surface of a patient. The portion of the device that captures images of the retina is shown transparently to show the sensor microarray and control board inside. That camera housing is connected by wire to the control unit in the background, which the eyecare professional uses to view images, change setting and view patient data. The small size of the device and its ability to be placed directly on the ocular surface makes it much more agile than traditional retinal imaging devices.