Research on the computational imaging method of full-visible-spectrum achromatic diffractive lens based on improved enumeration search algorithm

Compared to traditional optical lenses, diffractive lenses offer advantages such as compact size, lightweight, high integration, and flexible design, significantly enhancing the miniaturization, lightweight, and integration levels of imaging systems. However, diffractive lenses inherently suffer from significant chromatic aberrations, which severely limit their direct application in wide-band imaging fields. This paper focuses on the optimization design algorithms for achromatic diffractive lenses (ADL), analyzing the optimization efficiency and chromatic aberration reduction effects of current mainstream algorithms. It is demonstrated that when commonly used global search iterative algorithms are applied for the optimization design of large-aperture ADL, the parameter explosion leads to reduced optimization efficiency and suboptimal achromatic performance. Therefore, based on the design approach of computational optical imaging, this paper proposes an improved enumeration search algorithm based on concentric circle decomposition for optimizing the radial one-dimensional height parameters of ADL. Additionally, an image restoration method using the reweighted graph total variation (RGTV) prior and blind deconvolution algorithm is employed. Compared to other global optimization algorithms, this optimization algorithm not only achieves better results but also significantly reduces computational resource consumption. Furthermore, simulation results confirm that the optimized design exhibits better tolerance to height errors induced by manufacturing processes. The optimized ADL with a 5mm aperture and 50mm focal length, both continuous and discontinuous, were fabricated and tested for imaging. Experimental results validate that under the optimized design, the ADL achieve diffraction-limited imaging resolution, and combined with image restoration techniques, they can deliver clear object imaging across the entire visible spectrum.