Research On The Key Technology For Sequence Image Matching Aided Navigation System Based On High-dimensional Combined Feature

Image matching technology has important theoretical significance and huge development prospects in the field of autonomous navigation.In view of the shortcomings of existing image matching navigation technology,the key technology of integrated navigation system is thoroughly studied in this paper,with the background of sequence image matching aided inertial navigation.This article utilized image features to realize image navigation,and proposed an algorithm for constructing high-dimensional combined features based on stable branch points.The high-dimensional combined feature is presented in the form of intersecting lines pair,which has a certain geometric structure,easier to construct,and less susceptible to noise,at the same time combine the advantages that point features are relatively easy to extract and line features are stable and easy to match.On this basis,the article presented a sequence image matching algorithm based on high dimensional combined features,analyzed the geometric constraint relationships between features,and presented a method of eliminating the accumulative error through the absolute matching between reference image and real-time image at regular intervals,on the basis of the relative matching between adjacent frames in sequence images.Because of the uneven distribution of high-dimensional combined features and the poor robustness of the matching algorithm only using the geometric relations of features,this paper proposed a method improving high-dimensional combined features based on Delaunay triangulation to ameliorate the distribution of features and reduce redundant calculations.Meanwhile,in view of the shortage that geometric features are easily affected by image deformation,this paper presented the regional characteristics of high-dimensional combined features.On this basis,combined the geometric and regional characteristics of high-dimensional combined features together,the ten-dimensional feature descriptors of high-dimensional combined features are formed,which make the robustness of the algorithm is further enhanced.Finally,in order to verify the performance of the algorithm based on the high-dimensional combined features proposed in this paper,The output information of the image sensor,the altimeter and the inertial navigation system are fused together to form a integrated navigation system.At the same time,an improved Kalman filtering algorithm correcting delay according to inertial navigation information increment was studied,to achieve multi-information fusion,and the output of navigation system was optimized.Moreover the article designed and developed a visual integrated navigation simulation system and validated the effectiveness of the algorithm in this paper.