针对医院病房巡视机器人的工作环境复杂多变、障碍物形状各异的特点,提出了一种基于视觉-激光传感器的障碍物检测方法.通过视觉获得障碍物的大小、远近信息.结合激光传感器对环境的感知能力,构建虚拟障碍物,并形成避障策略关键字.通过查询避障策略表,实现病房巡视机器人在复杂环境下的安全避障功能.实验结果验证了该方法的有效性. Aiming at the complex and changeable working environment of the robot in the hospital ward and the different shapes of the obstacles, a detection method of the obstacle based on the vision-laser sensor is proposed. Obtain the size and distance information of the obstacle visually. Combining with the laser sensor pair Environment, building virtual obstacles and forming obstacle avoidance strategy keywords.An obstacle avoidance strategy table is queried to realize the safety obstacle avoidance function of ward-patrolling robot in complex environment.The experimental results verify the effectiveness of the method.