Defogging Algorithm for Aerial Image of UAV Can Resist Fog
The progress of UAV aerial photography technology promotes the increase of UAV aerial photography height and field of view.The improvement of the field of view and shooting height makes it easier to capture the fog image in a foggy environment.Most of the existing dehazing algorithms deal with the uniform fog image captured by the image sensor at a low angle of view,while the fog removal effect of the fog image captured by UAV at a high angle of view is poor.This paper proposes an unmanned aerial vehicle aerial image dehazing algorithm(MLD-Net)that can resist the influence of cloud fog.The uniform light module is introduced to adjust the uniformity of the features of the fog area of the image,enhance the detail loss caused by the local over-dark caused by the fog occlusion,and improve the image quality.The output image of the MLD-Net network presented in this paper outperforms the existing research results in terms of information entropy,average gradient and other evaluation indicators.The output image has rich detail features and is easier to be processed by advanced algorithms in the later stage.The proposed algorithm is applicable not only to the case of non-uniform fog,but also to the case of thin fog and thick fog.
万方数据
维普
