Infrared Dim and Small Target Sequence Dataset Generation Method Based on Generative Adversarial Networks

With the development of infrared technology, dim and small target detection plays a vital role in precision guidance applications. To address problems insufficient dataset coverage huge actual shooting costs methods, this paper proposes method for generating sequence datasets based on generative adversarial networks (GANs). Specifically, first, improved deep convolutional network (DCGAN) model is used to generate clear images sky background. Then, target–background are constructed using multi-scale feature extraction conditional networks. This fully considers characteristics background, which can achieve effective expansion image data provide test set recognition algorithm. In addition, classifier’s performance be by expanding training set, enhances accuracy effect learning. After experimental evaluation, generated similar real dataset, after with latest learning model.