A new approaches of Hydroxyl Tagging Velocimetry(HTV) in the application of supersonic combustion flow field signal extraction from images with strong background noise was proposed. These approaches are consisted of three steps: signal recognition through Hough transformation processing, image segmentation using maximum betweenduster variance algorithm based on Region of Interest (ROI), and tagging line extraction using a process combining skeleton extraction and direction template. Based on the above approaches, the image identification ability in the harsh combustion environment with strong background is improved, and the problem of insufficient accuracy in extracting hydroxyl effective signal is solved.