星载云检测算法的FPGA实现

对于光学遥感卫星而言，云的存在遮挡了地物，对卫星图像数据获取的质量造成了很大影响。对云进行实时有效的检测，可以把云图识别并分离出来，根据卫星的不同需求进行分类处理。一方面，为了减轻卫星数据传输通道的压力，云图可以不对地面传输；另一方面可以对云图进行高倍数压缩，由此节省出来的数据量可以给非云图像提供更高的压缩质量。提出的一种云检测FPGA实现方法，检测效果良好，厚云检测率达到97%，目前已成功应用于多颗卫星。

FPGA implementation of cloud detection algorithm used on satellites

For optical remote sensing satellites, the existence of clouds blocks ground objects, which has a great impact on the quality of satellite image data acquisition. Based on real time and effective cloud detection, cloud images can be identified and separated from original images, and then different processes can be conducted according to different needs of satellites. On the one hand, in order to reduce the stress on the satellite data transmission channel, cloud images do not need to be transmitted to the ground; on the other hand, cloud images can be compressed with a high compression ratio, and the amount of data saved can be employed to offer higher compression quality for non cloud images. This paper proposes an FPGA implementation method for cloud detection, which achieves a detection rate of 97% for thick clouds. Because of its good detection effect, the proposed FPGA implementation has been successfully deployed on a large number of satellites.