Research on Monocular Vision Distance Measurement Algorithm Based on Reference Target

Aiming at the problem that calculating the distance of the target to be measured relative to the camera for the case where the target has been detected in the intelligent video surveillance, this paper proposes a method for distance measurement only using a single target image without any internal camera parameters. This method firstly calibrates some distance sample points accurately through corner detection and localization, and then builds the mapping model between image ordinate pixel and actual distance through rational function fitting. Combining with artificial bee colony algorithm, the distance model is obtained after parameter optimization, so as to achieve the measurement of monocular image distance. The method requires only one image to be calibrated, and the effects of the imaging model, the imaging system error, and the lens distortion are not separately considered, but are implicitly solved in the regression fitting. The experimental result shows that the algorithm is effective and satisfies the real-time and accurate requirements of distance detection.