Processing Images Obtained from UAVs for Preserving Contours in Order to Classify Shapes

For most image processing algorithms, the scatter value is an important parameter. Estimating it as accurately as possible will allow obtaining optimal results considering both the amount of noise removed from the image and the quality of contour preservation. Estimating the dispersion by values lower than the correct one leads to the removal of a smaller amount of noise, and the estimation by higher values generates the blurring of the contours. In this article we propose an image processing architecture obtained from UAV based on wavelet decompositions for highlighting contours and recognizing/classifying objects.