The embodiment of the invention provides an image registration method combining target detection and semantic segmentation, and the method comprises the steps of obtaining a to-be-processed image fromthe sensor data, and intercepting a key region in the to-be-processed image through a real-time target detection algorithm; carrying out the scale invariant feature transformation on the key area toobtain the feature points in the key area; carrying out semantic segmentation on the key area based on the semantic information, classifying the feature points in the key area, defining an error function for each type of semantic information items, and obtaining a total error function by combining the error function in an original framework of the visual odometer; and carrying out optimization processing based on a likelihood domain model on the overall error function, and completing the image registration when the error is minimum. The extraction and matching range can be reduced by adoptinga key point extraction processing mode for the key area.