Invention:
This technology is a method of processing the interference fringes produced by the UA19-178 system. It employs a threshold-type operation at the incoherent light level so that only constructive interference is recorded, which greatly speeds up the data processing by reducing the amount of data to process in the first place. Processing occurs on the sensor to manipulate the raw signal into the desired form. The sensor’s circuit elements transform the data from each pixel without sending the signal off-chip, thus avoiding the bottleneck effect that occurs from serially transferring the data.
Background:
Event-based vision sensors offer a radical departure from traditional synchronous (frame-based) sensors. Rather than collecting data from an array of pixels at set times, these sensors provide a stream of asynchronous events corresponding to a specified change at a particular pixel address. This means that vision sensors can respond to changes ten times faster than traditional video cameras running at 60 hz while providing data of interest rather than redundant or stale data. However, challenges remain to processing the asynchronous output stream and calibrating pixels’ response within the extensive dynamic range of the sensor.
Threshold processing refers to the circuitry that processes the detectable signals from photo-sensitive elements and converts the signals to digital output that is interpreted by downstream algorithms. Concerns of precision and noise guide choice of circuit design, as well as cost.
Applications:
- Circuit modifications proposed in this project are tailored for applications involving white light interferometry imagers because they consider the specific computations useful to capturing depth of field and intensity data.
Advantages:
- The chief advantage of this modification would be processing speed in systems using white light interferometry imagers
