Abstract: The dynamic characteristics of the energy storage mechanism spring are closely related to the operation state of the circuit breaker. To address the problems of fast action time and difficulty in capturing kinetic parameters of the circuit breaker operating mechanism, a new method of detecting energy storage spring deformation parameters based on high-speed image sequence target tracking is proposed, in which a high-speed camera captures video images of the moment when the energy storage spring is released and a Lucas-Kanade optical flow algorithm for visual tracking is introduced to calculate parameters such as spring deformation speed and stroke. To solve the problems of spring deformation information extraction and real-time target tracking at the test site, an improved Lucas-Kanada optical flow acceleration algorithm is proposed to improve image recognition efficiency. The improved algorithm matches the image target at a speed of 4.6 ms/Frame in the field experiments for circuit breaker spring performance tests, demonstrating that the improved algorithm can meet the accuracy and real-time requirements for detecting the mechanical characteristic parameters of circuit breakers. The improved algorithm has unique advantages for specific high-speed continuous motion target detection, and the non-contact image tracking detection method has broad application prospects in circuit breaker status discrimination.