【摘要】 全景相机在机器人导航方面有着重要的应用。设计并实现了一套基于FPGA的双通道全景图像采集系统用于机器人导航。通过配置两个CMOS图像传感器的寄存器,实现了图像的同步采集。两个CMOS图像传感器的前端均采用性能相同的大视场鱼眼镜头,利用FPGA实现了全景图像的实时拼接算法,大大方便了后续的图像使用和处理。所得到的全景图像可通过USB3. 0实时传输到其他高性能处理器上。与传统的全景图像采集系统相比,不仅实现了实时全景图像处理,还具有低功耗、可扩展性强、使用方便等优点。更多还原

【Abstract】 Panoramic cameras are very useful in robot navigation. This paper designs and implements a panoramic image acquisition system based on FPGA and dual-fisheye lenses. By configuring the register of two CMOS sensors,the system achieves synchronization shooting. What’s more,a real-time image stitching algorithm is implemented by the FPGA,so that the output images are already stitched which greatly simplify the usage and processing in the subsequent progress. The images captured by this device can be transferred to other devices through the USB3. 0 in real time. Compared with the traditional panoramic image acquisition system,the device not only implements the real-time image processing,but also has advantages of low power consumption,high scalability and ease of use.更多还原