CN-120947639-B - Bionic polarized light navigation enhancement method and system for extreme weather

CN120947639BCN 120947639 BCN120947639 BCN 120947639BCN-120947639-B

Abstract

The invention discloses a bionic polarized light navigation enhancement method and a bionic polarized light navigation enhancement system for extreme weather, which comprise the steps of obtaining sky polarized original data and auxiliary navigation information, identifying invalid parts in the data according to failure criteria of underexposure, low polarization degree and local inconsistency, generating mask images for identifying valid and invalid areas, inputting the original data, the auxiliary information and the mask images into a pre-trained physical information neural network with internal polarized light field physical model constraints, carrying out predictive restoration on the invalid areas by the network by utilizing physical reasoning capacity of the network, combining the original data of the valid areas, outputting a predicted polarization mode graph, and calculating heading information of a carrier based on the polarized mode graph.

Inventors

HUANG RONGHUA
BAO YONGCHENG
HU YANG
YANG JINLEI

Assignees

江苏正方交通科技有限公司

Dates

Publication Date: 20260508
Application Date: 20250815

Claims (9)

1. The bionic polarized light navigation enhancement method for extreme weather is characterized by comprising the following steps of: acquiring sky polarization original data, synchronously acquiring auxiliary navigation information, identifying an invalid part in the sky polarization original data according to a preset failure criterion, and generating mask images of an effective area and an invalid area of identification data; Inputting the sky polarization original data, auxiliary navigation information and mask images into a pre-trained physical information neural network, carrying out information prediction on the invalid region based on a polarized light field physical model in the physical information neural network, and outputting a predicted polarization mode diagram by combining the original data of the effective region; Based on the predicted polarization mode diagram, calculating heading information of the carrier; Further comprises: And calculating and outputting the confidence coefficient of the course information of the calculated carrier by evaluating the coincidence degree of the predicted polarization mode diagram and the polarized light field physical model and the integral polarization intensity of the predicted polarization mode diagram.
2. The extreme weather oriented bionic polarized light navigation enhancement method according to claim 1, wherein the sky polarized raw data is a stokes vector image, the auxiliary navigation information comprises time, geographic position information and carrier posture information, and the failure criteria comprise underexposure criteria, polarization degree lower than a preset threshold criterion and data local inconsistency criteria.
3. The extreme weather oriented bionic polarized light navigation enhancement method of claim 1, wherein the physical information neural network employs an encoder-decoder structure and is provided with a jump connection that conveys multi-scale spatial features between the encoder and decoder.
4. The extreme weather oriented bionic polarized light navigation enhancement method of claim 1, wherein training of the physical information neural network is accomplished by minimizing a hybrid loss function configured to simultaneously cause the output of the physical information neural network during training to satisfy the following conditions: The first condition is that in a data effective area marked by a mask image, the output of the physical information neural network is consistent with the sky polarization original data; And the second condition is that in the whole mask image domain, the output of the physical information neural network accords with the rule defined by the polarized light field physical model.
5. The extreme weather oriented bionic polarized light navigation enhancement method of claim 4, wherein the polarized light field physical model is defined as a partial differential equation that satisfies the following physical effects simultaneously: the effect of spreading and smoothing polarized light in space; polarized light generates the effects of intensity attenuation and depolarization due to atmospheric scattering; the source effect of the directivity of the entire polarization mode, determined by the position of the sun.
6. The extreme weather oriented bionic polarized light navigation enhancement method of claim 5, further comprising: in the polarized light field physical model, taking the coefficient of the attenuation and depolarization effect and the intensity of the source effect as model-learnable parameters; and carrying out combination optimization on the learnable parameters and the weights of the physical information neural network in the training process.
7. An extreme weather-oriented bionic polarized light navigation enhancement system based on the extreme weather-oriented bionic polarized light navigation enhancement method according to any one of claims 1-6, characterized by comprising: The data acquisition and preprocessing module is configured to acquire sky polarization original data, synchronously acquire auxiliary navigation information, identify an invalid portion in the sky polarization original data according to a preset failure criterion, and generate mask images for identifying an effective area and an invalid area of the data; The polarization mode prediction repair module is configured to input the sky polarization original data, the auxiliary navigation information and the mask image into a pre-trained physical information neural network, predict the information of the invalid region based on a polarized light field physical model in the physical information neural network, and output a predicted polarization mode diagram in combination with the original data of the effective region; And the course information calculating module is configured to calculate the course information of the carrier based on the predicted polarization mode diagram.
8. A computer device comprising a memory and a processor, said memory storing a computer program, characterized in that the processor implements the steps of the method according to any one of claims 1-6 when executing said computer program.
9. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1-6.

Description

Bionic polarized light navigation enhancement method and system for extreme weather Technical Field The invention relates to the technical field of computer vision and autonomous navigation, in particular to a bionic polarized light navigation enhancement method and system for extreme weather. Background The bionic polarized light navigation technology is used as an autonomous navigation means for simulating living things such as desert ants and the like to position and orient by utilizing a sky polarized light mode, and has wide application prospect in the fields such as unmanned aerial vehicles, autonomous robots, vehicles and the like due to the characteristics of being passive, free of accumulated errors and high in precision. Specifically, the technology forms a stable all-sky polarized light distribution mode which takes the sun/anti-sun meridian as the symmetry axis by generating predictable Rayleigh scattering of sunlight by atmospheric molecules under the condition of sunny weather. The navigation system captures the mode through the polarization sensor, so that the solar azimuth angle can be accurately calculated, the carrier heading is further determined, and high-precision navigation is realized. However, due to the definition of the bionic polarized light navigation technology, the application scene of the technology is severely restricted in weather. In extreme weather conditions such as heavy fog, heavy rain, sand storm or heavy cloud, the atmosphere is filled with large-sized aerosols or water droplet particles with a size far greater than the wavelength of light, resulting in the transition of light transmission from rayleigh scattering dominant to complex mie scattering and multiple scattering dominant. The fundamental change of the physical process not only seriously damages the original stable polarization mode, but also causes the distortion of polarization information and the rapid reduction of signal to noise ratio, and even causes a depolarization area in which the polarization information completely disappears in a wide sky area. Existing enhancement methods focus on filtering or multi-source data fusion of signals, which simply treat signal degradation in extreme weather as noise superposition, and lack deep modeling of the complex scattering physical mechanism behind. Therefore, when the large-area information loss is faced, but not only noise interference is not only faced, the traditional method cannot repair and reconstruct the destroyed polarization mode fundamentally, so that the navigation precision is broken and cliff-like, even completely ineffective, and the application of the bionic polarized light navigation technology in all weather and high robustness is greatly limited. Disclosure of Invention This section is intended to outline some aspects of embodiments of the application and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section as well as in the description of the application and in the title of the application, which may not be used to limit the scope of the application. The present invention has been made in view of the above-described problems occurring in the prior art. Therefore, the invention provides a bionic polarized light navigation enhancement method facing extreme weather, which is used for solving the problems in the background technology. In order to solve the technical problems, the invention provides the following technical scheme: in a first aspect, the present invention provides a bionic polarized light navigation enhancement method for extreme weather, including: Acquiring sky polarization original data, synchronously acquiring auxiliary navigation information, identifying invalid parts in the sky polarization original data according to a preset failure criterion, and generating valid and invalid mask images of identification data; Inputting the sky polarization original data, auxiliary navigation information and mask images into a pre-trained physical information neural network, carrying out information prediction on the invalid region based on a polarized light field physical model in the physical information neural network, and outputting a predicted polarization mode diagram by combining the original data of the effective region; And calculating the course information of the carrier based on the predicted polarization mode diagram. The bionic polarized light navigation enhancement method for extreme weather is characterized in that the sky polarized original data are Stokes vector images, the auxiliary navigation information comprises time, geographical position information and carrier posture information, and the failure criteria comprise underexposure criteria, polarization degree lower than a preset threshold criterion and data local inconsistency criteria. As a preferable scheme of the bionic polarized light navigation enhancement method facing extreme weather, the physical inform