A Study of Imagery Translation Strategies in English and American Poetry Aided by Natural Language Processing Technology

British and American Imagist poetry has a profound influence on the entire Western literary world, and it is the poetry form of the British anti-Romantic literary movement. The language expression of the poetry is more implicit, and the language structure is concise and clear, with very strong modernist characteristics. In this paper, we design a strategy for translating English and American poetry imagery with the assistance of natural language processing technology, i.e., we construct a machine translation model based on Transformer’s chapter context validity recognition through the corpus to realize the accurate translation of English and American poetry imagery, and experimentally analyze the effect of the model. The method in this paper achieves the expected performance, with a maximum improvement of +1.99 BLEU compared to the sentence-level baseline model and a maximum improvement of +0.94 BLEU compared to the chapter-level baseline model, and achieves the optimum among a series of typical chapter-level translation models compared. Through the statistics of the deep meaning of imagery, it is known that the same deep meaning can be expressed by things of different meaning categories, and through the ratio of imagery type and frequency, it is known that poets will choose different imagery to express the same deep meaning. At the same time, in English and American poetry, a large number of rhetorical devices, such as borrowing and simile, are used. The deeper meanings of the imagery mostly reflect negative and painful emotions. The deeper meanings of many of the imagery are extremely rich, which reflects the polysemous nature of poetry. This paper lays the foundation for a better study of a chapter translation model that can fully perceive and efficiently utilize chapter context.