Laser Scanned Point Clouds for Selective Compression Masks

The patent replaces text recognition algorithms (software-based mechanical system) with laser scanner data processing (physical sensing system) to identify regions of interest. The laser scanner directly detects geometric features and distances, substituting the complex computational text recognition process with a more efficient physical measurement approach that naturally identifies traffic signs and other objects without requiring font, spacing, or rotation assumptions.

The patent applies differential compression where different compression factors are applied to different regions of the image based on their importance. Regions identified as containing traffic signs or other important objects through laser scanner data are assigned lower compression factors to preserve quality, while less important regions receive higher compression factors for maximum data reduction. This local quality approach ensures that critical information is preserved while achieving overall data size reduction.

The patent makes the region identification system universal by using laser scanner data that can detect any object with geometric features, not just text. The laser scanner's time-of-flight measurements and geometric analysis work equally well for traffic signs, trees, street lamps, buildings, and other objects without requiring specific assumptions about font, spacing, or rotation. This multi-functional approach allows the same system to identify diverse objects in various conditions.