Research On Image Optimization Method For Identifying Mechanical Characteristic Parameters Of Circuit Breaker

High voltage circuit breaker is an important equipment in power system and its operation reliability is vital to the protection and control of power grid.The mechanical characteristic parameters must be strictly tested before type test,factory inspection or transfer test,since it's the effective evaluation index for the performance of circuit breaker.Meanwhile,for the circuit breaker working in the transformer substation,mechanical characteristic parameter s also needs to be inspected regularly.The operation faults of circuit breaker can be prevented through abnormality of mechanical characteristic parameters and regular maintenance.Traditional testing methods which need to install sensors,such as grating sensor or sliding resistance sensor,have low test accuracy and can't test online.In view of the above shortcomings,a non-contact testing method based on computer vision technology is proposed: Capture high speed image sequences during the operation of circuit breaker and use feature part on the body as target.The normalized cross correlation(NCC)algorithm is combined with degraded pyramid search method and sector identification region of the following images is adjusted dynamically.Thus the matching algorithm is optimized.The optimized method in this paper is called normalized cross correlation-pyramid-sector(NCC-P-S)algorithm.In the process of tracking the target,make displacement of fixed part on the circuit breaker body as reference and use motion difference from adjacent frames to correct the vibration error.Moreover,multiple targets tracking are carried out to overcome the influence from surrounding environment.Finally,moving trail of the main shaft is obtained and mechanical characteristic parameters can be calculated by the relationship between main shaft and circuit breaker contact.The experiment proves: NCC-P-S algorithm has high matching speed and precision which can be well applied in this experiment.In addition,vibration error and environmental influence have been corrected by the correction identification method and thus the target coordinates are ensured accurate acquisition,so the test precision of mechanical characteristic parameters is improved.