An end-to-end Structure-strategy Defect Detection Network (SSDDNet) is proposed in this research to solve the problems in surface defect detection of metallic workpieces, such as complex defect morphologies, diverse scales, ambiguous boundaries, unstable label quality, and the high cost of pixel-level annotation. A multi-scale contextual aggregation module is structurally established to fuse dynamic convolution and multi-dilation information, while a boundary enhancement module is introduced to strengthen the modeling of ambiguous boundaries. A spatial label uncertainty modeling approach is strategically introduced to enable stable training under weakly annotated conditions. Experimental results show that: (1) SSDDNet achieves a 1.1 percentage points improvement in mean precision over the state-of-the-art MixSup model on the public KolektorSDD2 dataset without pixel-level annotations. (2) SSDDNet outperforms MixSup by 17.6 percentage points in mean accuracy and achieves approximately 15 percentage points improvement in SSH on the self-constructed industrial dataset BatteryBase, indicating the strong generalization capability of the proposed model. This research provides a novel approach for surface defect detection of metal workpieces.

In the process of UAV swarm road inspection, there are many problems such as difficulty in UAV route planning, unbalanced utilization of UAVs, and difficulty in determining the distributed airport site. In response, firstly, an inspection map is built to remove the redundant information irrelevant to road inspection. Secondly, route planning and airport site selection are unified in the framework of multi-objective optimization. Thirdly, the particle coding method, particle update rules and particle decoding method are proposed which combine route optimization and airport site selection. Fourthly, several evaluation indexes are proposed comprehensively to evaluate the effect of route planning and airport site selection. The experimental results show that: (1) By this method, the mileage of repeated part in optimized inspection route is less than 7% of the total mileage, and the UAV utilization balance rate is more than 75%. (2) After optimization, the reuse rate of distributed airports has been significantly improved. It shows that the method proposed in this research can plan the route for UAV road inspection task preferably, balance the utilization of UAV, and select a better location as the distributed airport site. It provides a firm foundation for autonomous UAV swarm road inspection system.