基于深度学习的天网图像质量评价研究

Research on Skynet Image Quality Evaluation Based on Deep Learning

天网图像质量评价从人工检查的主观评价法,到无参考图像质量评价方法,有了很大进步,但还存在准确度不高的问题。文章介绍了基于将深度学习的图像质量评价方法及其在天网图像质量轮巡系统中的应用。通过构建卷积神经网络模型,以从天网抓取的图片作为训练数据,并通过数据标定、数据预处理、模型训练等实验步骤,对图像清晰度、曝光和偏色三个方面进行评价,获取应用效果数据,验证模型的有效性。

The quality evaluation method of Skynet is not high,but there is still a great progress from the subjective evaluation method to the subjective evaluation method of image quality.This paper introduces the image quality evaluation method based on deep learning and its application in Skynet image quality inspection system.By constructing the convolution neural network model,taking the pictures captured from Skynet as the training data,and through the experimental steps of data calibration,data preprocessing and model training,the image definition,exposure and color deviation are evaluated to obtain the application effect data and verify the effectiveness of the model.