Insulators, pivotal to the integrity of railway catenaries, demand impeccable functioning to prevent system failures. Their consistent assessment is vital for railway safety. Current insulator evaluations in subways predominantly involve human intervention, a method fraught with inefficiencies, inaccuracies, and oversights, exacerbated by the complex backdrop of subway tunnels and minuscule defect dimensions. This study introduces an enhanced algorithm, anchored in the YOLOv5 framework, to refine insulator defect identification. Challenges in defect detection include limited, imbalanced data samples and adaptability. Addressing this, an accurate catenary model mirrors the subway line’s architecture, facilitating the creation of synthetic instances of both intact and impaired insulators. An atomization technique augments the dataset volume, fortifying the algorithm’s resilience in reduced visibility conditions, such as fog. Tackling sample equilibrium, the study introduces an equilibrium loss function, assigning disparate weights to various sample categories during training, thereby sharpening the algorithm’s focus on positive instances, particularly those that are challenging to discern, and rectifying the disproportion in sample categories. Incorporating lightweight structures like GhostNet and the Efficient Channel Attention Network (ECA-Net) channel attention scheme not only diminishes the network’s computational demands, thereby elevating the detection capabilities, but also minimizes superfluous data processing, enhancing the accuracy in identifying smaller targets. Empirical analyses indicate substantial model optimization: a size reduction to 60 pp of its original (from 15 MB to 9 MB), a near 1.4 pp increase in mean average precision (mAP) to 96.57%, and a tripling of the detection speed (from 30 to 90 FPS). Real-world image assessments further reveal a mAP improvement of approximately 2.5 pp (reaching 98.43%), confirming the model’s suitability for real-time applications.

A high-speed camera was applied to observe the shaking detachment tests with ‘Guiwei’ litchi clusters using 7 Hz, 9.5 Hz, and 12 Hz and the amplitudes of 40 mm, 60 mm, and 80 mm. The fruit’s detachment time and the number of fruit–fruit impacts were affected by the vibrational parameters and the fruit position (inside or outside of the cluster). A high and stable detachment speed and a low number of impacts could be achieved with the parameter combination of 9.5 Hz/80 mm. The cluster index, considering both the shaking position and the total fruit mass, was found to have a stronger correlation with the detachment time than the average fruit-to-shaker distance (distance from the shaker to the fruit). Thus, the detachment time can be estimated based on the cluster indexes. The fruit-to-shaker distance also affected the fruit-to-shaker impact. A fruit-to-shaker distance of 50~60 mm was recommended to avoid almost all the impacts, depending on the shaking parameters. The horizontal throw distance of the fruit was calculated based on the detachment velocity and angle, which can be used as the design basis for the catching devices. With a high enough drop height, a disc-shaped catching surface with a radius of 475 mm, 489 mm, 591 mm, 741 mm, 844 mm, and 832 mm, respectively, could collect 90% of all the fruit at six different vibrational parameter combinations. A closer catching surface to the fruit could effectively reduce the horizontal throw distance.