Abstract

This publication describes methods for detecting and mitigating image light-contamination when using active stereo cameras. Positioning an illuminator (or light projector) close to a camera underneath a cover glass of a user equipment (UE), such as a smartphone, can result in a direct optical-path reflection from the illuminator to the camera. If light “leaks” from the illuminator into one or both cameras, the emitted light contaminates the image(s) with extra light, which makes stereo matching difficult. To mitigate image light-contamination the UE leverages the symmetry of the cameras. First, the UE manufacturer calibrates both cameras to have consistent exposure with respect to each other. At the time of use, the UE captures one image from each camera with the illuminator enabled. The UE may perform this step when it detects a significant amount of energy (brightness) difference between the images. After the UE captures one image from each camera with the illuminator enabled, the UE performs a two-dimensional (2D) image (signal) processing. It is possible and, in some cases, prudent, the UE may directly perform statistical measurements in 2D signals. Nevertheless, performing statistical measurements of the 2D signals is computationally expensive. To simplify the computations, the UE converts the 2D signals into one-dimensional (1D) signals. After the UE converts the 2D signals into two 1D signals, the UE computes a statistical measure of the difference between the two 1D signals. After the UE has calculated the statistical measurements to determine whether image contamination has occurred, the UE may ask for user input when the distance is larger than a predefined threshold distance. Lastly, the UE performs stereo-matching using the images.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS