Spot Center Detection Algorithm For Continuous Frame Extraction In Natural Scene

Abstract:
In the process of axis calibration of the photoelectric pod, the spot emitted by the laser emitter needs to be used as the image reference center. Therefore, whether the position of the center of the laser spot illuminated by the laser emitter can be accurately detected is the key to the success of axis calibration. This paper presents a method to extract the spot center from the sampled continuous frames in the natural complex background. In the process of spot image sampling, the image jitter will be encountered. The image registration method is used to register the front and rear images; The method of frame difference is used to extract the background of the image; Using the method of maximum variance between classes to find the threshold of accurately extracting the light spot, and solving the maximum connected region to determine the light spot; Finally, the center of laser spot is determined by solving the gray center of gravity. This method has remarkable effect and strong adaptability. It is suitable for real-time axis calibration of photoelectric pod in natural scenes, and meets the needs of axis calibration.