航拍图像质量评价及其在图像增强中的应用

无人机在航拍过程中,由于受太阳光及大气环境、无人机姿态变化等因素干扰,导致采集图像出现失真,对后续信息处理造成不利影响。针对上述问题,提出了一种基于卷积神经网络的无参考航拍图像质量评价方法,并将其应用于航拍图像增强处理中。首先,设计了一种多层卷积神经网络。网络包括5层卷积层、5层池化层和3层全连接层;其次,进行卷积层和池化层多层堆叠学习图像特征信息。最后,将学习到的特征通过三层全连接层的回归与分类得到航拍图像质量分数。在自建的失真航拍图像库中进行实验表明,所提方法预测出的航拍图像质量分数与人眼视觉感知具有较高一致性,且在航拍图像增强领域应用效果较好。

Aerial Image Quality Evaluation and Application in Image Enhancement

When the Unmanned Aerial Vehicle(UAV) is taking pictures, it is easily to be affected by some factors such as sunlight, atmospheric environment, and posture adjustment of UAV, etc., which can lead to image distortion and influence the results of subsequent image information. In order to solve this problem, this paper proposed a non-reference aerial image quality assessment method based on Convolutional Neural Network(CNN) and used it in aerial image enhancement processing. First, a multi-layer CNN was designed, which includes five convolution layers, five pooling layers and three fully connected layers. Second, the image feature information was learned by the multi-layer stacking of the convolution layer and the pooling layer. Finally, the quality score was obtained by the regression and the classification of full connection layer. The experimental results on the self-built distortion aerial image database show that the predicted image quality score is highly consistent with the human visual perception, and it works well in the field of aerial image enhancement.