Joint first and second order total variation decomposition for remote sensing images destriping
Ayoub Boutemedjet,Sid Ahmed Hamadouche,Nabil Belghachem,Ayoub BoutemedjetSid Ahmed HamadoucheNabil Belghachema Laboratoire des Systèmes Lasers,Ecole Militaire Polytechnique,Algiers,Algeriab Laboratoire des Systèmes Optoélectroniques,Ecole Militaire Polytechnique,Algiers,AlgeriaAyoub Boutemedjet was born in Batna,Algeria,on 1989. He received the B.S.,degree in Ecole Militaire Polytechnique,Algiers,Algeria in 2012,and PhD,degree in electronics from the Beijing Institute of Technology,Beijing,China,in 2018. His research interests the calibration and correction of infrared focal plane arrays,remote sensing image denoising and terahertz detection and imaging.Sid Ahmed Hamadouche was born in Mascara,Algeria,on 1990. He received the M.S.,degree in electronics and telecommunication systems from Ecole Supérieure des Transmissions,Algiers,Algeria,in 2017. He is currently a PhD student at Ecole Militaire Polytechnique,Algiers,Algeria. His research interests the correction of fixed pattern noise in infrared images of focal plane arrays.Nabil Belghachem was born in Algiers,Algeria,on 1988. He received the M.S.,degree in optoelectronics in the Military University of Technology (MUT),Warsaw,Poland,in 2013,and PhD degree in optoelectronics from MUT in 2017. He has done many research studies and projects in the fields of passively q-switched microchip lasers and laser remote sensing devices. He is an author and coauthor of dozens of scientific and conference papers.
DOI: https://doi.org/10.1080/13682199.2024.2320491
2024-02-24
The Imaging Science Journal
Abstract:Stripe noise remains a significant source of errors and image quality degradation in remote sensing systems. A prominent approach for tackling this problem is the first-order Total Variation (TV) regularization, which has a proven efficiency in dealing with stripe noise. Unfortunately, denoised images, in this case, may suffer from texture loss and staircase artefacts around smooth areas. In this paper, a novel image decomposition scheme is proposed to tackle these drawbacks. This decomposition is based on the use of directional first- and second-order TV regularizers that are employed to separate the clear image from the stripe component while considering the directionality and smoothness of the latter. The proposed model is solved using a Chambolle-based algorithm and its performance is compared to traditional destriping methods using different noise structures and intensities. The results have shown comparable performance to the existing state-of-the-art methods with some improvements in structure preservation and noise cancellation.
imaging science & photographic technology