Abstract: A new technique for flare angle detection based on moving image identification is proposed. It contains several technical measures as follows. The visual sign was designed for recognition and the synchronization between image acquisition and the speed of elevator governor was realized. The mathematical morphology was used for image pre-processing and the visual sign was recognized by using filtering and shape parameter extraction technology. Then, the flare angle of flyweight was calculated and the method of least square was used for quadratic curve fitting between flare angle and speed. The testing results indicate that the correct rate of detection can be stabilized at 97% or more.