This study proposes an improved rail bearing platform measurement method based on a binocular line-structured light sensor to alleviate the problems pertaining to the existing rail bearing platform detection methods. During the measurement, the relative position of the two cameras in the binocular sensor and rail bearing platform remains unchanged. Scanning of the rail bearing platform is performed by controlling the line laser movement using the driving device built in the sensor to obtain the point cloud data, which improves the data acquisition effect and measurement accuracy of the embedded sleeve position of the rail bearing platform. Concerning point cloud data processing, a method for point cloud data segmentation and extraction based on the Spearman rank correlation coefficient is proposed. Automatic, noncontact measurement of five key geometric parameters of rail bearing platform is conducted. The experimental results show that the accuracy of the proposed measurement method is up to 0.3 mm, which can fulfill the requirements of track plate detection.