Snapshot high dynamic range imaging based on adaptive frequency stripe masks

Abstract

A novel snapshot method based on adaptive frequency stripe masks is proposed to realize high dynamic range (HDR) imaging for highly reflective surfaces. The high refresh rate of the digital micromirror device is used to load modulation masks, and superimposed modulation of the target scene is performed to achieve snapshot HDR imaging by using a single superimposed image. Adaptive frequency stripe masks are generated based on the energy spectrum of the scene to enhance the lateral resolution of the demodulated sub-images. Then the sub-images are separated through a frequency domain reconstruction algorithm, and the final HDR image is obtained by using a multi-exposure image fusion technique. Experimental results demonstrate that the proposed method produces ghost-free HDR images with enhanced lateral resolution and contrast while effectively avoiding overexposure and underexposure. This approach provides a valuable solution for visual imaging of highly reflective surfaces.