Abstract: The research status of the key technologies of multi-rotor detection unmanned aerial vehicles(UAVs) in mine dangerous areas at home and abroad is reviewed from three aspects: autonomous positioning and navigation technology, autonomous obstacle avoidance technology and multi-sensor information fusion technology. Autonomous positioning and navigation technology enables robots to move autonomously in unknown environments without human intervention. Combined navigation technology, 3D environment map building technology, deeply optimized trajectory planning algorithms and simultaneous positioning and map building technology based on semantic and deep learning are suitable for mine conditions with complex and unstable mine environmental conditions where disaster information evolving over time.The obstacle avoidance method based on multi-sensor information fusion can ensure that the multi-rotor detection UAV can perceive obstacle information to the maximum extent under different environmental conditions. The sensor fusion architecture based on autonomous positioning and obstacle avoidance technology need to adopt a distributed structure to make the mine multi-rotor detection UAV system have high reliability and fault tolerance.The problems of multi-rotor detection UVAs are analyzed from both software and hardware aspects. The problems include that the universality of the fusion model and algorithm cannot be guaranteed,the fusion system's fault tolerance or robustness needs to be improved, the lack of processing hardware to adapt to a variety of complex fusion algorithms, and the low degree of integration, high power consumption and large size of multi-sensor.The development trend of key technologies of multi-rotor detection UAVs in mine dangerous areas is prospected. ① Fusion algorithm optimization: it is crucial to maximize the optimization of fusion algorithm, improve system reliability and stability and ensure stable and efficient data processing. ② Application of artificial intelligence technology: improving the deep learning ability of multi-rotor detection UAVs and expanding the detection range of mine dangerous areas by applying intelligent technologies such as machine learning and adaptive technology. ③ Development of processing hardware that can adapt to multiple complex fusion algorithms: the conditions of mine dangerous area are extremely complex, and it is difficult to achieve simultaneous collection and processing of underground multi-source information without the processing hardware that can adapt to the deep fusion of multiple algorithms. Therefore, the information processing ability of multi-rotor detection UAVs can be improved by developing processing hardware with strong adaptability. ④ Development of convenient hardware fusion system: developing a fusion system based on the deep integration of multiple sensors could further enhance the detection capabilities of multi-rotor detection UAVs.