With the increasing penetration of power electronics at the receiving end of line-commutated converter based high voltage direct current (LCC-HVDC) systems, the electrical signals at the point of common coupling show time-varying behavior, leading to growing wideband resonance risks in the system. To address the poor adaptability and narrow measurement bandwidth of traditional impedance identification methods under the gray-box modeling context of the LCC-HVDC receiving-end grid, this paper proposes a non-intrusive dynamic wideband impedance identification method under multiple operation conditions. First, an equivalent wideband impedance model of the LCC-HVDC receiving-end system is established. Based on the port signals at the grid connection point, synchrosqueezed wavelet transform is applied to extract wideband dynamic features. Then, an improved Kalman filtering algorithm is adopted to achieve rapid fusion of feature information and time-frequency domain estimation of impedance, enabling dynamic tracking of the grid wideband impedance without external disturbance. Finally, the wideband impedance characteristics and resonance risks of the LCC-HVDC system are identified. The effectiveness of the proposed method in impedance characterization and wideband resonance risk assessment is verified using generalized Nyquist curves and time-domain simulations.