一种基于改进太阳子午线拟合方法的仿生偏振光定向系统

为设计一种自主性强、精度高、抗电磁干扰能力强的定向系统,以昆虫获取大气偏振模式的感知机理为基础,设计了一种新型的成像式仿生偏振光定向系统.该系统使用具有微偏振阵列结构的CMOS相机作为主要的偏振信息采集设备,使用DSP/ARM架构的NVIDIA JETSON TX2嵌入式处理器进行信息解算,具有体积小、运算能力强等优点.通过研究大气偏振模式中偏振度和偏振角图像的对称性,针对已有航向角解算方法在太阳子午线提取精度方面的不足,设计了一种对称轴粗提取之后通过连续旋转法精确提取太阳子午线的算法,最终实现定向精度0.3335°,在最小二乘对称轴粗提取法的基础上定向精度提升了19.05％.

A Bionic Polarized Light Orientation System Based on Improved Solar Meridian Fitting Method

In order to design a directional system with strong autonomy, high precision and strong anti-interference ability, a new imaging bionic polarized light orientation system based on the perception mechanism of insects to obtain atmospheric polarization mode is designed. The imaging bionic polarized light orientation system uses a CMOS camera with a micro-polarization array structure as the main polarization information acquisition device, and uses a high-performance DSP/ARM architecture based NVIDIA JETSON TX2 embedded processor for information calculation. It has the advantages of small size and strong computing performance. By studying the symmetry of the polarization degree and the polarization angle image in the atmospheric polarization mode, to address the low accuracy of the existing imaging polarized light orientation system''s heading angle acquisition method in extracting the solar meridian, an accurate solar meridian extraction algorithm by continuous rotation method after a rough extraction of the symmetry axis has been designed in this paper. This algorithm has an orientation accuracy of 0.3335°, and on the basis of the least squares symmetry axis coarse extraction method, the orientation accuracy is improved by 19.05%.