(Universidade do Estado do Amazonas, 2020-12-01) Alves, Cássio Antonio Tavares; Oliveira, Jozias Parente de; Del Rio, Daniel Guzmán; Santos, Karlo Homero Ferreira dos
The study in the area of Digital Signal Processing is one of the ones that most attract
researchers due to the fact that many products or services today require image processing, since
human beings are essentially visual. The present work has as its theme the Development of a
System for the Automatic Detection and Removal of Shadows in Images of Indoor
Environments. Where the objective is to deal with the subject of digital image processing in
order to develop a system that is able to detect and remove shadows automatically, which is
configured as a pre-processing step in computer vision tasks. The present work presents a
shadow detection system based on the K-Nearest Neighbors (KNN) method, a non-parametric
statistical method, widely used in image segmentation tasks. The approach used is very
interesting since the results obtained show a rate of up to 85% of correctness in detecting
shadow zones in the images.
