 Pig posture detection has become increasingly important for the welfare and productivity of captive pigs. Traditional methods are time-consuming and expensive, while deep learning techniques offer a more efficient and accurate solution. The review discusses the development of various data collection methods and subtasks, as well as the application of deep learning models in pig posture detection. Limitations of current methods in future directions for research are also discussed. This article was offered by Zhichin, Zhixin Liu, and Haiyan Wang.