The characteristics of rope vibration of the winding mine hoist are one of the main evaluation indices for evaluating the characteristics of the engineering mechanics and the safety coefficient of the hoist system. For this purpose, research was conducted on the vision-based measurement method for catenary transverse vibration. Firstly, a high-speed camera was utilized to capture grayscale images of the catenary transverse vibration. Secondly, the region of interest (ROI) was selected, and the ROI was binarized based on the wavelet edge extraction algorithm. Thirdly, in the ROI, the centroid position of the rope was obtained by calculating the matrix, which was described by the binary, and later, the transverse vibration signal of the catenary was obtained. Finally, the empirical mode decomposition (EMD) method was used to remove the periodic excitation caused by the displacement of the catenary winding. The experimental results show that the proposed method quickly and accurately comprehends the noncontact measurement of the catenary transverse vibration, and this method could also provide the technical support necessary for the measurement and evaluation of the safety parameters of the hoisting system.