Blurred Image Detection In Drone Embedded System
Ratiba Gueraichi, Amina Serir · 2020
This paper deals with the detection of blurred images that may eventually be captured by a drone. The embedded system should be able to measure the amount of blur affecting the images in order to decide whether to acquire the scene again or not. For this purpose, we have developed a simple model based on Discrete Cosine Transform (DCT) associated to Support Vector Machine Classifier SVM, to classify images into three categories and thus detect strongly, moderately and slightly blurred images. The proposed system has been tested on 550 images captured by a drone. The obtained results are very conclusive.