PROJECT TITLE :

Single Image Defogging Based on Illumination Decomposition for Visual Maritime Surveillance

ABSTRACT:

The elimination of single picture fog is critical for surveillance applications, and recently, numerous defogging methods were presented. Aside from depth information in a foggy scene, atmospheric aerosol model, which has a greater impact on illumination in foggy scenes than in an otherwise clear environment, also affects the scattering qualities. However, newer defogging methods mix haze with fog, and therefore don't take into account the scattering qualities of the fog completely. To remove fog effects from surveillance photos, these methods are insufficient. A single picture defogging technique has been proposed for visual maritime surveillance in this paper as a result. Modeling the scattering of light in the glow-shaped lighting, a fog image is proposed. To remove the glow effect from the airlight radiance, an illumination decomposition algorithm is presented to recover a fog layer, in which the objects at the infinite distance have uniform brightness. It's then used to limit the transmission map to an acceptable range by using an estimation of the nonlocal haze-lines prior. As a final result, the suggested illumination compensation technique allows defogging images to retain the original image's natural lighting information. The visual marine surveillance is further supported by a fog image dataset. When comparing subjective and objective evaluation criteria, the experimental results show that the suggested method outperforms current methods. As a result, the proposed technology may remove fog while preserving its natural appearance.


Did you like this research project?

To get this research project Guidelines, Training and Code... Click Here


PROJECT TITLE : Iterative Refinement for Multi-source Visual Domain Adaptation ABSTRACT: One of the most difficult aspects of multi-source domain adaptation is figuring out how to minimize the differences in domains that exist
PROJECT TITLE : Learning Versatile Convolution Filters for Efficient Visual Recognition ABSTRACT: This article presents versatile filters that can be used to construct efficient convolutional neural networks, which are widely
PROJECT TITLE : Deep Visual Odometry with Adaptive Memory ABSTRACT: A novel deep visual odometry (VO) method that takes into account global information by selecting memory and refining poses is presented here. The currently available
PROJECT TITLE : Iterative Refinement for Multi-source Visual Domain Adaptation ABSTRACT: One of the most difficult aspects of multi-source domain adaptation is figuring out how to minimize the differences in domains that exist
PROJECT TITLE : A Review of Single-Source Deep Unsupervised Visual Domain Adaptation ABSTRACT: Deep neural networks have been shown to perform exceptionally well across a broad spectrum of benchmark vision tasks as a result

Ready to Complete Your Academic MTech Project Work In Affordable Price ?

Project Enquiry