No abstract

No abstract

No abstract

Detection in large-scale scenes is a challenging problem due to small objects and extreme scale variation. It is essential to focus on the image regions of small objects. In this paper, we propose a novel Adaptive Zoom (AdaZoom) network as a selective magnifier with flexible shape and focal length to adaptively zoom the focus regions for object detection. Based on policy gradient, we construct a reinforcement learning framework for focus region generation, with the reward formulated by object distributions. The scales and aspect ratios of the generated regions are adaptive to the scales and distribution of objects inside. We apply variable magnification according to the scale of the region for adaptive multi-scale detection. We further propose collaborative training to complementarily promote the performance of AdaZoom and detection network. To validate the effectiveness, we conduct extensive experiments on VisDrone2019, UAVDT and DOTA datasets. The experiments show AdaZoom brings consistent and significant improvement over different detection networks, achieving state-of-the-art performance on these datasets, especially outperforming the existing methods by AP of 4.64% on Vis-Drone2019.

The paper studies the problem of how to recognize aircraft targets staying at the airport from satellite image database of huge amount in an acceptable processing speed. Without adopting the current method recognizing the target picture by picture from the image database, the paper combines image retrieval with target recognition, and firstly uses image retrieval technology to pick out those images containing airport target from the satellite image database, then utilizes target recognition technology to recognize the aircraft targets staying at the airport from those images. Some new methods or thoughts have been put forward about the airport image retrieval, segmentation and recognition for aircraft targets, and a retrieval and recognition system for aircraft targets staying at the airport in the database of satellite images has been studied and designed. Finally many experiments has been designed and carried out, and the experimental results demonstrate that these methods are feasible and effective.