To address the demand for effective lateral tracking of the Pantograph–Catenary contact point on Electrified Highways, this paper proposes a machine vision-based detection method for determining the contact point position. The method first enhances image acquisition accuracy by optimizing the image sensor layout and removing lens distortion using Zhang’s calibration method. Then, it extracts Catenary edge information using LAB color space thresholding and the Canny edge detection algorithm. The initial Catenary position is determined using perspective transformation and histogram statistics, followed by precise Catenary fitting using a sliding window and the least squares method. A detection model incorporating the relative coordinates among vehicle, Pantograph, and Catenary is established to achieve high-precision contact point detection. Field experiment results show that the method achieves a mean absolute error of 0.0062 m under sufficient lighting, although performance deteriorates under low-light conditions. This study is of great significance for improving the operational safety and reliability of Electrified Highway Pantograph–Catenary systems.