Weeds is an important factor leading to crop yield reduction, how to identify accurately is the premise and basis of weed management. With the progress of computer and information technology, the combination of machine vision and image processing has become the mainstream method of weed detection and recognition. From the point of view of image preprocessing, segmentation, feature extraction and classification, the current research progress of field weed recognition at home and abroad, as well as the advantages and disadvantages of various segmentation, extraction and recognition methods were introduced in detail. In addition, the effects of light environment, the occlusion and overlap of leaves, and the optimization of classifiers in weed detection in the field were analyzed and discussed. Finally, the suggestions and prospects for the current research trend of weed identification were put forward. 