Endoscopic image luminance enhancement based on the inverse square law for illuminance and retinex.

Abstract

In a single-port robotic system where the 3D endoscope possesses two bending segments, only point light sources can be integrated at the tip due to space limitations. However, point light sources usually provide non-uniform illumination, causing the endoscopic images to appear bright in the centre and dark near the corners. Based on the inverse square law for illuminance, an initial luminance weighting is first proposed to increase the image luminance uniformity. Then, a saturation-based model is proposed to finalise the luminance weighting to avoid overexposure and colour discrepancy, while the single-scale retinex (SSR) scheme is employed for noise control. Via qualitative and quantitative comparisons, the proposed method performs effectively in enhancing the luminance and uniformity of endoscopic images, in terms of both visual perception and objective assessment. The proposed method can effectively reduce the image degradation caused by point light sources.