Abstract: Mine coal and rock identification is one of the key problems restricting intelligent coal mining. Affected by the geological conditions and complex working conditions of coal mines, many coal rock identification methods cannot meet the real time, intelligent and efficient application requirements of industrialization due to their technical limitations. On the basis of summarizing the evolution history of coal and rock identification technology in mines, this paper proposes that the coal and rock identification technology has entered a new stage of simultaneous mining, autonomous and dynamic detection. Also, it proposes a theoretical and technical framework of non contact intelligent detection of coal rock interface suitable for intelligent fully mechanized mining face. The framework includes four technical connotations of electromagnetic wave energy focusing, posture perception, support posture adaptive control, and coal rock interface coordinate conversion, focuses five key technologies of emission air coupled radar antenna, intelligent adjustment support, coal rock interface identification and tracking technology methods, interface position information extraction technology, and system pose parameter matrix. The basic principle of the coal rock interface dynamic detection system based on the air coupled radar antenna is systematically explained. Firstly, an air coupled shielded radar antenna is developed to realize the focused emission of electromagnetic wave energy and non contact detection of coal rock interface. The aerial height of the antenna is 30 cm. Secondly, the smart antenna support is developed, combined with the support attitude adaptive control technology, based on the airborne distance detected by the radar data (the distance between the antenna and the coal surface), to realize the self adaptive adjustment of the radar antenna attitude with the fluctuation of the working face. Thirdly, the coal rock interface identification and tracking method under multiple interference sources is investigated for identifying the primary effective wave reflection points of the air coal interface and coal rock interface accurately and obtaining dynamic and high precision coal rock interface information. Finally, the system pose parameter matrix is created, combined with the inertial navigation unit of the coal machine and the dynamic pose parameters of the intelligent support, the spatial relationship model of the geodetic coordinate system, the shearer coordinate system and the antenna coordinate system are established. The contour curve of the coal rock interface is created. The information above is transmitted to the coal machine control system through the network protocol, so as to realize the self adaptive height adjustment control of the shearer cutting drum. The experimental results show that the average error of identifying the coal rock interface position using the coal rock recognition system is ±1.2 cm, and the average error percentage is 8.6%. The dynamic detection of the coal rock interface in the mine is realized, which provides a new theoretical and technical support for the unmanned operation of the intelligent mining face.