Liquid column manometers are widely used in fluid mechanics experiments, so the design and implementation of a remote experimental setup for performing experiments on fluid mechanics, the on-line reading of liquid column manometer is necessary. Firstly, through comparing computer vision reading with manual reading, the former accuracy is much higher in the ideal case. Secondly, computer vision based reading scheme is confirmed according to the experimental progress. Then, the imaging geometric model of manometer is approximated by pin-hole imaging model. While the vertical relationship between camera viewpoint and liquid level is changed, for reducing the imaging error, view fixed reading with software algorithm is discussed to correct the identified water head. Lastly, according to the proportional relations between invariable inside diameter of manometer and water head image, the upper boundary line coordinate value is determined. The experimental results show that the method is effective and feasible.