The modernization construction continues to develop rapidly, and the elevator industry in China has achieved rapid growth. Currently, the annual production and installation of elevators in China rank first in the world. Chinese elevators have the characteristics of high frequency of use, long operating time, and complex operating environments. In recent years, there have been multiple accidents in China caused by the failure of elevator brakes, resulting in casualties. The safety of elevators has received increasing attention. The elevator brake is the most critical component to ensure the safe and reliable operation of the elevator. The main function of elevator brake is to stop the car to the set position. In addition, the brake is also provided with functions such as car ascending protection and car unexpected movement protection. Elevator safety accidents caused by brake failure are usually more serious. Therefore, this study proposes a visual-based monitoring method for elevator drum brakes. This method uses a camera to capture the motion image of the elevator brake wheel and processes the image through Fourier spectral analysis. Based on the correspondence between spectral images and texture images, the actual action state of the brake wheel is determined, and the braking condition is analyzed by comparing it with historical data. This monitoring method only requires exposing the brake disc, achieving non-destructive monitoring of elevator brakes, which is of great significance for improving elevator safety levels, preventing and reducing elevator accidents.