Point Image Restoration for Spherical Aberration Correction

The patent pre-calculates and stores restoration filters for multiple F numbers before actual image processing. When an image is captured, the system simply selects the pre-computed filter corresponding to the current F number, avoiding the need to perform complex restoration filter calculations in real-time. This preliminary preparation significantly reduces the computational burden during actual image restoration while maintaining the ability to correct spherical aberration effectively.

The patent dynamically selects different restoration filters based on the actual F number used during image capture. Instead of using a single fixed restoration filter, the system adapts by choosing from multiple pre-computed filters that correspond to different F numbers. This dynamic selection ensures optimal restoration performance for each specific imaging condition while avoiding the complexity of real-time filter adaptation.

The patent replaces the need for expensive mechanical/optical solutions (high-precision imaging optical systems designed to minimize spherical aberration) with a computational approach. Instead of relying on complex optical design and expensive lenses, the system uses software-based restoration filters that mathematically correct spherical aberration effects. This substitution of mechanical/optical complexity with computational processing achieves similar image quality improvement at lower system cost.

The patent creates a computational model (restoration filter) that replicates the effect of perfect optical performance. Rather than building expensive optical systems that physically prevent spherical aberration, the system captures images with simpler optics and then applies restoration filters that copy or recreate the appearance of images that would have been captured by a perfect optical system. This allows achieving high image quality without the need for expensive optical hardware.