|
The bionic polarization navigation sensor has important research and application value in the field of modern navigation. In this study, a new algorithm for calculating the solar meridian azimuth from imaging polarization navigation sensors is proposed. By analyzing the sky polarization distribution model,there is only one line with a polarization angle of 90° in the projection plane of the sensor, and the zenith point of the sky must be on this line, with which a novel method for extracting the solar meridian in the two-dimensional projection plane is proposed. A polarization measurement model is built based on the Stokes vector method. Gaussian smoothing filtering is performed on the polarization angle image to conduct image noise suppression. The preliminary positioning of the solar meridian is solved by a dual-threshold recursive method.Then, the accuracy of the meridian detection is promoted to the sub-pixel level by interpolation to detect the exact pixel points on the solar meridian. Finally, the least-squares fitting of the precisely positioned pixels are utilized to obtain an accurate azimuth angle.An outdoor verification experiment is carried out with an imaging bionic polarization navigation sensor with a high-precision turntable system.The experimental results show that the static repeatability measurement accuracy of the algorithm is 0.0554° (σ) . The dynamic maximum error is less than 0.95° . This indicates the effectiveness and feasibility of the direction angle calculation method. The method can provide accurate and no cumulative error heading information for the movement of the carrier and provides important technical support for the field of navigation.
|
