针对海上搜救的背景,研究了基于全景视觉的移动机器人进行海上救助的关键技术。首先,提出了基于颜色模型的图像分割方法,运用区域增长的方法处理分割图像并进行目标定位。然后,配合声纳传感器对目标进行准确的抓取。最后,通过实验检验图像分割和目标定位的精度,验证了该方法运用到海上救助系统的可行性。 In view of the background of maritime search and rescue, the key technologies of salvage for mobile robots based on panoramic vision are studied. First of all, a method of image segmentation based on color model is proposed. The method of region growth is used to process the segmentation image and locate the target. Then, with the sonar sensor to accurately capture the target. Finally, the feasibility of applying this method to the marine salvage system is verified by experimentally testing the accuracy of image segmentation and target localization.