基于CUDA的航天遥感图像实时快视系统的研究

Research on Aerospace Remote Sensing Images Real-time Quick-view System Based on CUDA

随着实践中航天下行数据速率和类型的剧增,地面遥感图像实时快视系统需要有更强的数据处理能力和更好的显示效果。对于大批量的数据,采用传统的中央处理器(CPU)进行数据处理耗时较长,而且滞后严重。为了解决这个问题,文中提出向实时快视系统中引入CUDA平台的方案,利用GPU强大的通用计算能力加速图像数据的解析和显示。利用这项技术,可以获得数倍于传统CPU处理方案的效率,并且可以优化显示效果,而且随着硬件设备的升级,效率还可以得到进一步的提升。

With the rapid growth of space downlink data rate and type,ground sensing imaging real-time quick-view system requires faster processing power and better display effect.For large amounts of image data,the traditional method of data process conducted by CPU will cost too much time and have a great hysteresis.To solve the problem,present a method of introducing the CUDA platform to the real-time quick-view system,accelerating the process and display of image data by means of great computing capacity of GPU.With this method,the system can achieve a processing speed several times faster than traditional processing method by CPU,which makes the real-time image display better.In addition to this,it can get higher efficiency with the promotion of hardware.