Automatic video stream images evaluation and preprocessing for machine vision tasks

Chickrin Dmitry, Malyugina Alexandra, Derjavin Dmitry, Egorchev Anton

The quality of the input image stream is one of the typical problems in computer vision problems. In this paper, we consider an algorithm for preliminary filtering of video stream images, based on the evaluation and consecutive correction of the qualitative characteristics of images. At the evaluation stage, the algorithm uses global image properties such as contrast, noise level, sharpness, and also the presence of overlighting. For the stage of processing of images based on the evaluation of these properties, the algorithm uses image processing techniques such as histogram equalization, noise reduction and sharpness correction. The proposed algorithm is intended for use in various problems of machine vision.

Keywords:image preprocessing, image filtering, denoising, unsharp mask, histogram correction.