In this paper, an algorithm to remove noise particles such as snow and rain that may occur in natural
environments is proposed. Analyze the effects of de-noised images with the proposed algorithm can have
on the object detection network for an autonomous vehicle. The research process for this will be divided
into three stages. The first stage is the research preparation stage, and the hardware and data sets are
built. When constructing a data set, it is linked to the algorithm evaluation method. In the second
step, a development algorithm is proposed. To this end, approaches in the field of noise removal are
analyzed and an approach appropriate for the research purpose is selected. And, analyze representative
algorithms of the chosen approach and study image processing techniques essential for noise removal.
Through this, it understands the image processing techniques essential for noise removal and identifies
the constraints that occur when removing noise. And, by devising a method to overcome the constraints,
a noise removal algorithm would be proposed. In the third step, the proposed algorithm is evaluated and
the effect of the proposed algorithm on the autonomous driving object detection network is analyzed.
To this end, an algorithm to be compared is selected by understanding the principles of several recently
announced algorithms, and considering the characteristics and performance of the algorithms. To verify
the effect on autonomous driving object detection, the object detection algorithm is analyzed for the
effect on object detection by targeting the algorithm selected for each de-noise algorithm approach.