Research on the Intelligent Oracle Bone Script Recognition Model Based on OTSU and Improved YOLOv8

As the oldest known mature writing system in China, the identification and study of oracle bone scripts are of profound significance to understanding the origin of Chinese culture and the history of world culture. In view of the limitations of traditional oracle bone inscription recognition methods in the face of original oracle bone inscription images, this study proposed an intelligent recognition model based on OTSU Otsu method and improved YOLOv8 model to improve the recognition accuracy and efficiency of oracle bone inscription images. Firstly, an efficient oracle image preprocessing model is established by using OCR technology to remove the coding interference in the image, and using OTSU Otsu method to binarize the image, combined with bilateral filtering technology to reduce the noise. Secondly, the pre-processed images were input into the improved YOLOv8 model added to Slim Neck structure for training. By adjusting the model parameters, the performance of small target detection was significantly improved, and the accurate segmentation of oracle images was realized. Finally, combined with the image preprocessing and segmentation model, this study carried out text recognition on the images in the test set, and pre-trained the extended dataset with the YOLOv8-cls model, which enhanced the model's recognition ability of oracle bone characters, realized accurate recognition of oracle bone bones, and output the final result with high confidence. The intelligent recognition model in this study has shown significant performance improvement in image preprocessing, object detection, small target optimization and text recognition, etc. It provides a new solution for Oracle's digitalization and automatic recognition, and has important academic value and application potential for promoting the in-depth research of Oracle.