A method of UAV visible light remote sensing image registration based on eigenvector technique

As the advancement of unmanned aerial vehicle technology and deep learning technology, many deep learning technologies have been applied in the field of unmanned aerial vehicle image research. To improve the registration precision and efficiency of unmanned aerial vehicle visible light remote sensing images, this study uses the idea of eigenvectors to build an unmanned aerial vehicle visible light remote sensing image registration model. Firstly, the traditional registration model is optimized by multi-branch attention mechanism, and then the final remote sensing image registration model was constructed by combining convolutional neural network and mask prediction module. Testing the performance of the model, and the experiment findings expresses that the registration error of the final model is about 0.05, the registration precision is more than 0.95, and the registration time is within 2s, so it has a good performance of image registration. In conclusion, the image registration model designed by this research institute can improve the existing problems in the current registration field and provide new ideas for improving the image registration effect.