Abstract: To address the low efficiency of traditional contact-based measurement for critical clearances in circuit breaker mechanisms and the difficulty of stable boundary identification in complex structural images due to metallic reflections, background interference, and non-uniform gray-level variations, a visual measurement method for circuit breaker mechanism clearance based on adaptive gradient edge detection and gradient-weighted B-spline contour fitting is proposed. The proposed method constructs local gradient statistical features to adaptively adjust the edge detection threshold, thereby improving the extraction capability of critical boundaries under complex imaging conditions. On this basis, a gradient-weighted constraint is introduced into the B-spline fitting of edge point sets to achieve stable representation of the structural boundaries on both sides of the clearance, and the clearance size is then calculated in combination with calibration parameters. Experimental results show that the proposed method outperforms comparative methods in terms of boundary integrity, edge continuity, and contour tracking ability, and can achieve stable measurement of critical clearances in circuit breaker mechanisms. The method provides technical support for non-contact detection and condition assessment of related structural parameters.