Aiming at the difficulty of infrared image segmentation of buried landmine targets under complex background conditions, a region of interest (ROI) selection method based on morphology and clustering algorithm is proposed using the similarity of landmine shape characteristics and multiple landmine targets in the minefield.After eliminating the noise of the original image and suppressing the background by differential morphological filtering, the area where the target is located is reduced.Then the similar features of multiple targets in a certain area are used to cluster the suspicious areas, further reducing the target areas. The threshold segmentation is carried out respectively.Finally, the recognition is completed according to the relevant features of the target.The processing results of measured images show that this method has good segmentation effect and high positioning accuracy for buried multi-mine targets.In addition, the calculation speed of the algorithm is fast, which can meet the actual demand of mine detection.