Camera Lens Dust Detection and Dust Removal for Mobile Robots in Dusty Fields

摘要

In this paper, a new method of detection of the dust particle in camera lens placed on a moving robot is implemented. Then further for correction of recorded/live data a dust removal technique using adaptive tiling-based selective dust area coverage scheme has been implemented. An artifact such as dust is a main problem for camera lenses, which are used in many fields for example in photogrammetry, robotics inspection, geospatial data capture etc… The goal is to enhance the dust detection and correct the image data by comparing consecutive frames captured by moving robots and cleaning it by using the technique specified in the proposed system without much affecting the original data frame. Finally, our simulation results demonstrates that the proposed method was able to remove the dust particles with an accuracy of 90-92%. This accuracy without affecting the actual data is quite significant.