The Outsight 3D Semantic Camera simultaneously works on a broad spectral band, including visible and SWIR.
An easier way to visualise this is to start with how cameras perceive reality.
What we humans can perceive, is in fact only a tiny fraction of the electromagnetic spectrum:
The addition of Red, Green and Blue bands allows to define the broader range of colors that create the image.
The reflected wavelengths from objects, depending on their chemical structure characteristics, is what make us perceive their color:
We perceive the tree as "green" because it mostly reflects the green portion of the spectrum.
To encode the color of an image, each pixel contains the values of the amount of light reflected from each RGB band:
The combined amount of light from each RGB band creates the image.
And the relative intensities per wavelength defines what is called the Spectral Signature:
So how does this apply to the non-visible parts of the spectrum? In a very similar way, Outsight's 3D Semantic Camera can perceive both the visible and the Infrared part of the spectrum (SWIR for Short-Wave Infra-Red): the reflected light is not only originated by the ambient light but by a broadband laser that generates the light pulses.
Because the light that bounces back from objects is originated by the 3D Semantic Camera, it performs Active Sensing. And because it runs a continuous broadband of wavelengths simultaneously, it performs Active Hyperspectral Sensing.
Why is this relevant? Well, it happens that when looking at the SWIR part of the spectrum, the Spectral Signature of emblematic materials can be identified, even making out the difference between Ice and Snow:
Even if the broadband Laser generates a continuous signal over a wide bandwidth, depending on the application there is no need to process all this data at once.
Instead, a limited number of bands, eventually of different width, can be used, generating a multi-spectral response:
The perception of the visible spectrum is then supplemented by an active SWIR response:
Multi-spectral Remote Sensing is not a new subject, it has been widely used in Satellite Surveillance, Agriculture and Mining for years, in most cases by using passive 2D hyper-spectral cameras.
Outsight's approach is unique due to the combination of several factors:
- A novel broadband laser (patent-pending) generates the broad spectrum light, making it an Active Hyperspectral solution, that can be focused on the object of interest in order to concentrate the energy, hundreds of meters away. When required, the super-continuum data can be discretised into a Multi-spectral signal, making an even more efficient and simple device.
- The whole solution is integrated in a single small device, that spans from the visible spectrum to SWIR, in order to provide a comprehensive perception of the environment.
- The processing is done in real-time, including the multiple wavelengths' perception.
- The data output is a 3D multi-physics image.
- The same data structure also provides point-wise velocity, segmentation and classification, that combined with the Material ID, brings Situation Awareness to a whole new dimension.
For instance, detecting Black Ice when driving is good, but understanding also that it is on the road (and not in the sidewalk...), and more precisely on your lane, is much better. This is the difference between a Smart sensor and a Fully Aware one, between a system prone to false positives and a robust solution.
- Because we've built the solution with scalability in mind, it's an affordable one.
- For the same reason, its low power consumption is compatible with the constraints of challenging applications such as automotive.
The Outsight 3D Semantic Camera unveils the invisible information of the world. In a single device. In real-time.
This Full Reality Perception is a key enabler to fulfil our mission of bringing Full Situation Awareness to Smart Machines.