Depth Resolution via Analog Blur Restoration in ToF Systems

The patent applies preliminary degradation to the light reception data before depth estimation. By intentionally blurring the light reception image based on a known degradation characteristic (such as point spread function) before sampling, the system prepares the data in advance to enable super-resolution reconstruction. This preliminary action allows the restoration unit to later recover depth information with resolution exceeding the sampling period limitation, thus resolving the contradiction between measurement range and depth resolution.

The patent introduces an additional processing dimension by applying degradation and restoration operations in the spatial domain before depth calculation. Instead of directly sampling the light reception data, the system transforms the problem by convolving with a degradation kernel (point spread function) and then restoring using inverse filtering or iterative optimization. This dimensional transformation enables the system to achieve high depth resolution while maintaining wide measurement range, as the restoration process can recover fine depth details without being constrained by the original sampling period.

The patent changes the parameters of the light reception data by applying controlled analog degradation (blurring) with specific degradation characteristics such as point spread function models. By adjusting degradation parameters (blur radius, kernel type) and restoration parameters (regularization strength, iteration count), the system optimizes the balance between depth resolution improvement and processing complexity. The degradation characteristic is designed to match the optical system's actual point spread function, enabling effective restoration while keeping computational load manageable through parameter optimization.