Chinese natural scene speech recognition system based on OpenCV

The invention provides a Chinese natural scene speech recognition system based on OpenCV, and the system comprises the steps: carrying out the image collection of a natural scene through visual equipment, and obtaining a natural scene image; analyzing and processing the natural scene image by using OpenCV to obtain a Chinese annotation text of the natural scene image; processing the Chinese annotation text by using large-scale Chinese corpora, extracting a plurality of natural scene keywords and combining the natural scene keywords into a natural scene keyword text; and performing speech synthesis on the natural scene keyword text to generate natural scene description speech. According to the method, specific scene information such as different behaviors, equipment and environment information in the natural scene can be labeled, textualized analysis is carried out through keywords, specific textualized content is generated, texts can be converted into voices through the textualized content in a voice conversi