LiDAR Signal Processing Memory Segmentation for Resolution

The patent divides the histogram memory into multiple banks (first bank, second bank, third bank) that can be independently accessed and updated. This segmentation allows the system to maintain sufficient memory capacity for accurate TOF measurements while reducing the physical size of each individual memory bank, thereby resolving the contradiction between high resolution requirements and measurement performance maintenance.

The patent introduces a bank selection dimension to the memory architecture, transitioning from a single large memory space to multiple smaller memory banks organized in three dimensions (bank index, bin index, and counter). This dimensional reorganization enables the system to achieve the required measurement precision with smaller individual memory components, thus improving pixel resolution without sacrificing measurement performance.

The patent implements preliminary action by pre-organizing the histogram memory into multiple banks with specific addressing schemes before signal processing begins. The bank selection logic and multi-dimensional addressing structure are established in advance, enabling efficient signal processing despite the reduced memory size. This preliminary organization allows the system to maintain high productivity in TOF measurements while using compact memory structures.

The patent applies local quality by assigning different functions to different memory banks. The first bank stores historical data, the second bank stores current measurement data, and the third bank stores reference data. This functional differentiation allows each local memory segment to be optimized for its specific purpose, maintaining overall signal processing capability while reducing total device size through efficient memory utilization.