3D imaging is being used by vision and robotic systems to recover 3D scene geometry, which then use these images to support technologies such as autonomous transportation, augmented reality, and robot navigation. Time-of-flight (ToF) 3D cameras are bringing about a revolution in this space, because of their low-cost, compact form factors, and low computational complexity. They are now becoming more and more ubiquitous - from vehicular LiDARs to consumer grade devices like iPad, premium phones, AR/VR headsets, etc. These cameras work on the principle of emitting modulated light, and then, inferring the reflected light from the scene to capture the depth of the objects and image them.

However, they encounter a major hurdle - when multiple ToF cameras try to picture the same scene, they shall receive the light from each other as well, and hence, they suffer from huge depth errors. In this project, we want to understand the interference of light signals and accurately recover the 3D scene information. More specifically, we draw parallels between the image capture process of cameras, and the communication of wireless RF systems. Interference is a well studied problem in the wireless space, and we claim that we can leverage techniques from the wireless space, for countering the multi-camera interference problem.

Due to their low cost, compact form factor and low computational complexity; ToF based active 3D are now becoming more and more ubiquitous - even making their way into consumer grade devices like mobile phones, tablets, AR/VR headsets, etc. These cameras consist of a programmable light source that emits modulated (temporally coded) light, which travels to the scene of interest and gets reflected. This reflected light is received by the sensor (typically, co-located with the transmitter) and the scene depth is captured by comparing the reflected light with the emitted light. These cameras suffer from major interference, however, in the presence of multiple 3D ToF cameras. The following figure demonstrates the error in recovered depths when multiple ToF cameras image a particular scene. The red light is the received reflection intensity, while the true depth can be recovered when the camera recovers the depth from the true reflected light (in green).