基于注意力机制的高分辨率遥感图像目标检测

Object Detection of High-resolution Remote Sensing Images Based on Attention Mechanism

高分辨率遥感图像目标检测是计算机视觉实践课程中的一个创新实验项目。针对高分辨率遥感图像目标检测过程存在的问题,在主干网络的残差基本单元中引入注意力机制,设计基于注意力机制的目标检测模型。利用模型分别检测DIOR数据集及DOTA数据集,并对检测结果进行了对比分析。结果表明,该检测模型实现了对图像中不同目标较为精确的分类和定位功能,泛化能力较强。

High-resolution remote sensing images object detection is an innovative experimental project in the computer vision practical courses. Aiming at the problems in the object detection process of high-resolution remote sensing images, an attention mechanism was introduced into the residual basic unit of the backbone network, and an object detection model based on the attention mechanism was designed. The model was used to detect the DIOR data set and the DOTA data set, and the detection results were compared and analyzed. The results show that the detection model achieves more accurate classification and positioning functions for different objects in the image, and has a strong generalization ability.