The invention relates to a method and an apparatus for detecting traffic signs (1). In a method according to the invention for detecting traffic signs, an area surrounding the vehicle is recorded with the aid of a camera. The image data from the camera are analysed, wherein at least one image area (10) which contains a potential traffic sign is determined. The image area is submitted to a first classification unit which identifies the potential traffic sign on the basis of at least one class-specific feature as a traffic sign (1) of this particular class. A brightness or colour intensity analysis of the surface of the traffic sign (1) along radially running scanning beams (5) is then carried out starting from a defined starting point. Potential contour points (6) of an information-carrying part (3) of the traffic sign (1) are determined from the brightness or colour intensity analysis. An area of the surface of the traffic sign (1) is extracted as the information-carrying part (3) of the traffic sign (1) that includes all potential contour points (6) determined. The extracted area is passed to a second classification unit for semantic interpretation. The method according to the invention is distinguished by rapid carrying-out of the method on the basis of the brightness or colour intensity analysis along scanning beams (5), which, in contrast to two-dimensional edge-finding operators, for example the Sobel or Canny operator, which are otherwise used in traffic sign detection, imposes considerably lower demands on the control unit of a vehicle.