Scene matching suitability analyzing method based on multiple-feature integrating visual attention model

Abstract

The invention provides a scene matching suitability analyzing method based on a multiple-feature integrating visual attention model. The scene matching suitability analyzing method based on the multiple-feature integrating visual attention model can effectively improve reliability of abstracting a suitability zone and accuracy in scene matching. Compared with a traditional method, high efficiency and reliability of a visual attention sensing circumstance are fully utilized and SURF non-variable features are added into a traditional visual attention model. The multiple-feature integrating visual attention model is established after integrating cross-scale feature images and merging with features such as colors, strength and directions. A zone which obviously abstracts real-time maps serves as a suitability zone based on the model. Reliability of suitability analysis can be effectively improved and the scene matching suitability analyzing method based on the multiple-feature integrating visual attention model has important significance in application of scene matching practical engineering.

Description

technical field [0001] The invention relates to a scene matching adaptability analysis method based on a multi-feature integrated visual attention model (Area Suitability Analysis in Scene Matching based on Multi-Feature Fusion Visual Attention Model, MFF-VAM ASA), which realizes scene matching of drones The function of adaptive robust analysis in visual navigation can effectively improve the reliability of adaptive region extraction and the accuracy of scene matching. Background technique [0002] In the field of UAV scene matching visual navigation, adaptability analysis technology is crucial to navigation accuracy and performance. The selection of the adaptation area directly affects the reliability and effectiveness of the navigation system, and is the primary problem that must be solved in the scene matching visual navigation system. It has important theoretical significance and application value to study robust and reliable scene matching adaptability analysis methods...

AI Technical Summary

Problems solved by technology

Since the assumptions of the signal correlation model often cannot fit the actual situation, and the required parameters are often difficult to obtain, the adaptability analysis method based on image signal correlation calculation is not reliable

Similarly, the adaptability analysis method based on comprehensive feature evaluation generally determines the fit area by constructing a statistical model between the image feature index and the evaluation index, which requires a large number of experimental samples, and it is difficult to give general information about various ground objects and imaging conditions. Appropriate threshold

Moreover, the main factors affecting the adaptability include scene characteristics, imaging differences between the reference map and the real-time map, and matching performance requirements, etc., and the mutual constraints and interdependence of each factor lead to conflicts in the calculation results of the scene matching area. contradiction

Therefore, there are some deficiencies in the existing matching area selection methods.

Images