To address the challenges of substation instrument appearance defect detection, including insufficient semantic representation of dial damage under non-frontal viewpoints, complex background interference caused by dew and water-vapor pseudo-textures, and unstable localization of small-scale defects under long-distance imaging, this paper proposes a DLMFR-YOLO model based on YOLO11n. First, a Bidirectional Multi-scale Feature Amalgamation Propagation Network (BiMFAPNet) is constructed and embedded into the neck to realize bidirectional interaction and detail compensation among multi-scale features, thereby enhancing the semantic and structural representation of oblique-view and small-target defects. Second, a Star Attention Feature Extraction Module (SAFE) is designed, which leverages star operations and Context Anchor Attention to highlight key regions such as dial cracks and casing damage in both spatial and channel dimensions, while effectively suppressing interference from complex backgrounds and dew/water-vapor pseudo-textures. Third, WIoUv3 is adopted as the bounding box regression loss to improve the stability and accuracy of small-target localization through a dynamic quality weighting mechanism. On this basis, a lightweight StarNet backbone and a Lightweight Shared Convolutional Detection Head (LSCD) are introduced to compress the model parameters and computational cost. Finally, a logical distillation framework based on Bridging Cross-task Protocol Inconsistency Knowledge Distillation (BCKD) is constructed, where the pre-lightweight and post-lightweight models serve as teacher and student networks, respectively, to compensate for the performance degradation caused by lightweighting. Experimental results on a self-built substation instrument appearance defect dataset show that, compared with YOLO11n, DLMFR-YOLO improves mAP@0.5 by 3.2 % while reducing the number of parameters by 31.4% and FLOPs by 15.9%. This demonstrates that the proposed model achieves an effective balance between detection accuracy and model lightweighting, and exhibits favorable advantages for practical engineering deployment.