Quan Qi’s research while affiliated with Communication University of China and other places

Plain Language Summary Since the clear trace indicates the ionospheric electron density background and the Spread‐F indicates disturbances/irregularities, the prediction of both ionograms with/without Spread‐F is important to the research and application. There are currently no well‐established neural networks specifically designed to predict the extensive information encompassed within ionograms, particularly the intricate characteristics of Spread‐F. In this paper, we approach the short‐term prediction (next 30 min) of ionograms with/without Spread‐F. To achieve this, we employ the ConvGRU network to generate ionograms blurred but still captured the primary spatio‐temporal features of various types of Spread‐F (FSF/RSF/MSF/SSF) in ionogram sequences, as well as features without Spread‐F. Subsequently, we refine these rough images by EDSR network to obtain clear and detailed predictions. Our treatment to blurred prediction figures and our focus on the Spread‐F key area are innovative. Through our work, the auto‐prediction for ionosonde observation benefits ionosphere research and monitoring.

Plain Language Summary The study of ionospheric Spread‐F (SF) is of great scientific significance and practical value, particularly concerning the correlation between low‐latitude ionospheric scintillation and SF. Our research aims to develop an intelligent SF detection and classification method based on deep learning with high accuracy. It can greatly assist in auto‐analyzing ionogram images rather than manually. While deep learning models (particularly DCNNs) came to be used for SF detection, most of them were incapable of distinguishing SF types, such as FSF/RSF/MSF/BSF/SSF. Moreover, these models were not built on a large data set. Considering the strong regional and temporal variation features of SF occurrences and patterns, samples must cover various situations. The number of ionograms produced by ionosonde at Hainan station from 2002 to 2016 is nearly 520,000. We have manually classified these images into five categories (no SF and 4 types). On this basis, we did this research. The deep learning models in our paper perform well, with most of them being able to extract the features of different types precisely. Our research provides a handy tool for automatically classifying FSF/RSF/MSF/SSF based on image features only, without considering the ionogram data format or the auto‐models in ionosonde.

Interaction is the most important factor that may much influence gaming experience. Given that Augmented Reality presentations enable the more creation ways of user interactions that expand traditional HCI from 2D to 3D spaces, this offer an ideal opportunity for game developers to design more playful and interesting AR games that embed various interaction forms. In this paper, we build an augmented reality game creative platform, which allows game developers to integrate augmented reality features, spatial human computer interaction mechanism and other advanced interaction techniques into the designed game. In this system, we implement a adopted augmented reality function to help the developers to directly superimpose the visual 3d objects over the live real video image and then port human motion interactive trigger module to this system based on kinect, in which a up-down hierarchy figure description algorithm is proposed to build a common virtual 3d human model prototype that could allow developer ignore the difference of various players' shape and make the players easily to use their actions to interact with the 3d virtual objects to get more unique game experience.