The key problem in footbridge health monitoring and condition assessment is to accurately estimate time-varying modal parameters such as damping ratio and natural frequency of the bridge based on measured response signals. Thus， a new set of solutions is proposed. First of all， based on the HS optical flow method of segmented optical flow method for bridge vibration to the displacement response， Then， by redefining the objective function， the adaptive genetic algorithm was used to optimize the variational mode decomposition （VMD）， and then each displacement time history was decomposed， Hilbert transform （HT） curve was used to fit the single mode signal obtained from VMD decomposition to get the instantaneous frequency and damping ratio.Through the pedestrian bridge model， this solution is used to test the structural responses of controlled frequency and free walking performed by different pedestrians are tested， multiple sets of effective structural displacement responses are obtained and it is compared with the laser displacement meter， and the instantaneous frequency and damping ratio of pedestrian structure interaction obtained by VMD-HT decomposition are compared with the 941B sensor for verification. The results show that the visual non-contact measurement method adopting the segmented optical flow method shows high measurement accuracy for the structure response， and the error is 0.85% compared with the laser displacement meter. According to the measured signal， the VMD-HT is able to effectively estimate the time-varying modal parameters of the structure. By pedestrian-structure interaction， the natural frequency of the structure decreases with the increase of the number of people and the displacement， the maximum reduction is 14.12%， and the damping ratio increases with the increase of the number of people and the displacement， the maximum increase is 398.33%. Controlled frequency walking has a greater impact on time-varying modal parameters than free walking.