Remote sensing image text generation and optimization method based on self-reinforcement learning

Abstract

The invention provides a remote sensing image text generation and optimization method based on self-reinforcement learning. The remote sensing image text generation and optimization method based on self-reinforcement learning comprises the following steps: S1, extracting remote sensing image semantic understanding feature; S2, obtaining a training set, pre-training the text generation model, and extracting text generation model parameters; and S3, inputting the extracted feature vectors, a prior text library, pre-trained text generation model parameters and task requirements of a user into a remote sensing image text generation network, and restoring image feature information represented by the extracted feature vectors into text description through a deep learning natural language processing technology. The text is generated by adopting the self-reinforcement learning remote sensing image text generation algorithm based on the strategy gradient algorithm, the training effect of the remote sensing image generation model is improved, parameters are promoted to converge towards expected values, and the accuracy of generation description is improved.

Description

technical field [0001] The invention relates to the technical field of remote sensing image text generation, in particular to a remote sensing image text generation and optimization method based on self-reinforcement learning. Background technique [0002] The rapid progress of modern aerospace remote sensing technology has made remote sensing satellites more mature, and the amount of remote sensing data acquisition has doubled. However, at present, people's reasoning and understanding of aerospace remote sensing images is mainly based on detection and classification, and there is still a large gap between the results and high-level information. Therefore, in the face of such a large amount of remote sensing image data, it is urgent to have the interpretation ability that matches the acquisition speed of remote sensing images. How to mine and extract high-value information from the vast remote sensing images has become the direction of further exploration and research in the...

AI Technical Summary

Problems solved by technology

[0005] (1) The evaluation indicators used in the training stage and the verification stage of the remote sensing image text generation technology do not match

In the training phase, the cross-entropy loss is commonly used as the loss function training model parameter, but the verification phase uses BLEU (Bilingual Evaluation Understudy, bilingual evaluation method), ROUGE (Recall-Oriented Understudy for Gisting Evaluation, similarity measurement method based on recall rate), CIDEr (Consensus-based Image Description Evaluation, Consensus Image Description Evaluation), SPICE (Semantic Propositional Image Caption Evaluation, Semantic Propositional Image Caption Evaluation) and other evaluation indicators, resulting in descriptive sentences generated by the model and training samples may have semantic structure difference, which makes it difficult for sentences with lower loss values ​​during training to achieve corresponding effects during testing

[0006] (2) The remote sensing image text generation model cannot perform error correction

However, its model tends to converge to a local minimum, and takes longer to converge. Especially for remote sensing images, the model is more likely to fall into a local optimum because of its inconspicuous features, and the quality of the generated description is not good.

Images