A Vignetting Correction Algorithm for Bright-Field Microscopic Images of Activated Sludge

Vignetting correction is an essential part of pre-processing to address decreasing illumination towards the border of a microscopic image. Image processing and analysis is a potential tool for monitoring and prediction of activated sludge wastewater treatment plant. Microscopic images of activated sludge require anti-vignetting or vignetting correction algorithm for accuracy of subsequent image analysis procedures. In this paper, we proposed a vignetting correction procedure based on Gaussian modeling. The Gaussian model is estimated by using Otsu threshold and iterative evaluation of the model. The advantage of the proposed algorithm is that the vignetting model estimated for one image is found to be valid for all other images irrespective of illumination of microscope. Once the model is calibrated, the correction becomes simple addition, making the proposed technique time-efficient compared to state-of-the-art procedures. The assessment was done subjectively and by using segmentation. The proposed procedure performed better than polynomial approximation and comparable to Leong's algorithm.