In view of the low efficiency of manual inspection of oxygen cylinder distance, early warning of oxygen cylinder and acety-lene cylinder in the maintenance area of smart mine, a safety distance detection method for oxygen cylinder and acetylene cylinder un-der monocular video was proposed. Based on the improved DeepLabv3+ semantic segmentation models, this method segments the ima-ges of oxygen cylinder and acetylene cylinder, calculates the monitor distances according to the oxygen cylinder and acetylene cylinder after calibration and finally constructs a real-time safety warning algorithm to handle the alarm of oxygen cylinder and acetylene cylinder in smart mines. The results show that when the actual distance between oxygen cylinder and acetylene cylinder is 350 cm, 450 cm and 550 cm, the average errors of the target algorithm test results is 2. 37％, 4. 93％ and 5. 18％, respectively. The practical application of this method can improve the automatic monitoring efficiency of the distance between oxygen cylinder and acetylene cylinder under monocular vision,reduces labor costs and avoid safety accidents.
维普
万方数据