Year: 2022
Type: article
Abstract: With significant advancements in computer vision technology using deep learning algorithms, object detection has shown terrific performance in bright and clear targets. However, due to the low signal to noise ratio (SNR), the challenges of detec ting and segmenting objects in dim and dark targets continue to affect computer vision applications in dark and visually polluted conditions. Dim object detection using infrared imaging based on a deep learning algorithm, Convolutional Neural Network(CNN) for IR images, isproposed to counter this challenge. By studying and analyzing a series of CNN algorithms, the author presents an application based on Mask R CNN that is better in precision and speed to detect and segment target objects in infrared images dim and dark targets.
Authors: Weldekiros Misgana Desalegne, Rui Qing Wu
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7217544213
