Image defogging method based on online knowledge distillation

• Main Authors: CUI ZHIGAO, SU YANZHAO, ZHONG XIAO, LI QINGHUI, LI AIHUA, ZHANG WEI, LAN YUNWEI, WANG NIAN
• Format: Patent
• Language: Chinese; English
• Published: 31.05.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses an image defogging method based on online knowledge distillation. The method comprises the following steps: 1, acquiring a training set image; 2, establishing an image defogging network teacher model; 3, extracting features of the foggy training image; 4, establishing a total loss function; 5, training the image defogging network teacher model by the foggy training image; and 6, defogging a single image by using the trained image defogging network teacher model. According to the method, feature extraction is carried out through the subject network model, two defogged images are generated through the student branch network, feature aggregation is carried out through the feature aggregation network model, then knowledge distillation is carried out on the subject network model and the student branch network by utilizing the aggregation feature graph, the image defogging effect is improved, and network parameters are reduced. 本发明公开了一种基于线上知识蒸馏的图像去雾方法，包括步骤：一、训练集图像的获取；二、图像去雾网络教师模型的建立；三、有雾训练图像