Abstract: Based on analysis of principle and characteristics of mine flood perception methods such as water quality monitoring method, water inflow monitoring method, water temperature monitoring method, temperature monitoring method, humidity monitoring method, water level monitoring method, resistivity monitoring method, stress monitoring method, microseismic monitoring method and hydrological drilling method, mine flood perception method based on image monitoring and image-based big data mine flood perception and water source determination method were proposed. Conclusions were got as following: ① The water quality monitoring method can not only perceive flood, but also determine water source that causes flood, which has high perception accuracy of mine flood caused by surface water and goaf water. ② Time-based water inflow monitoring method has high accuracy, but complex deployment and poor real-time performance. The water inflow monitoring method based on flow velocity has advantages of simple operation and good real-time performance, but large measurement error. ③ The water temperature monitoring method can not only perceive mine flood, but also determine water source that causes flood, but it is not applicable to the flood perception that difference between mine flood water source temperature and normal mine water inflow temperature is small. ④ The temperature monitoring method has advantages of simplicity and convenience, but it is affected by mine fire such as coal spontaneous combustion, explosion of gas and coal dust, surface air temperature, mine ventilation quantity, underground equipments starting and stopping, and number of underground workers. The temperature monitoring method is not applicable to the flood perception that difference between mine flood water source temperature and normal mine water inflow temperature is small. ⑤ The humidity monitoring method has advantages of simplicity and convenience, but it is affected by humidity and temperature of surface air, mine ventilation quantity, and mine fire such as coal spontaneous combustion. ⑥ Change of surface water and groundwater source can be timely grasped by the water level monitoring method, but location of goaf water needs to be proved. ⑦ The resistivity monitoring method has advantages of fast response and high sensitivity, but accuracy is greatly affected by mining environment, and electrode arrangement is difficult. ⑧ The stress monitoring method and the microseismic monitoring method have advantages of good real-time performance, but due to impact of coal and gas outburst and rock burst pressure, the methods need to be used together with other mine flood perception methods. ⑨ The hydrological drilling method has advantage of large amount of information, but it needs to be used together with other mine flood perception methods. ⑩ The mine flood perception method based on image monitoring has advantages of non-contact, real-time and rapid, wide monitoring range, simple deployment and installation, low cost and convenient use and maintenance. Image-based big data mine flood perception and water source determination method simultaneously monitors mine water, water conducted channel and water source, which can not only perceive mine flood, but also determine water source that causes flood, and has advantage of high reliability.