Robust VSLAM algorithm based on multi-source information composite constraint model and heterogeneous kernel function optimization

Puntos clave

• The enhanced VSLAM algorithm achieves superior accuracy in dynamic environments with diverse input types.
• Achieving an optimization performance increase of 30% over traditional methods, the algorithm efficiently integrates multiple data sources.
• Assessment using a composite constraint model effectively reduces error margin in position estimation.
• This approach may enable better real-time navigation applications, although further validation is needed.