This project aims to simulate various visual distortions to improve the robustness and reliability of Monocular SLAM (Simultaneous Localization and Mapping) systems. SLAM is a technology that allows robots and autonomous systems to understand their surroundings using only a single camera. However, in the real world, cameras often operate under different environmental conditions, leading to visual distortions.
This project simulates common visual distortions, such as Raindrop, Mud, Noise, Motion Blur, and Lens Blur, to test how SLAM algorithms perform under these conditions.
Below are example images for each category of visual distortion. These images represent the effects of each distortion type and are used to test how SLAM algorithms perform under such conditions.
Original
Images in this category:
Raindrop20
Raindrop35
Raindrop50
Images in this category:
Mud130
Mud180
Mud230
Images in this category:
Noise2000
Noise3000
Noise4000
Images in this category:
MotionBlur15
MotionBlur20
MotionBlur25
Images in this category:
LensBlur52
LensBlur63
LensBlur74
The images presented above represent examples of various types of visual distortions. These images have been generated as part of this project to simulate real-world environmental conditions that could affect SLAM algorithms. The purpose of this simulation is to evaluate how well SLAM systems perform under different types of distortions such as raindrops, mud, noise, motion blur, and lens blur.
These images are intended as sample data to test and improve the robustness of monocular SLAM algorithms under challenging visual conditions.