This paper presents an enhanced version of YOLOv8 specifically designed for detecting damage in overhead power lines. The enhancements include an adaptive threshold mechanism, a novel convolution method called GSCov, and a lightweight network structure called Slim Neck. The adaptive threshold mechanism dynamically adjusts the detection threshold based on the brightness and contrast of the input image, improving the model's robustness. GSCov balances the model's running speed and accuracy by introducing a flexible way of information interaction through grouping and random permutation. Slim Neck reduces the model's complexity and computational load while maintaining good performance. Experimental results on the 'Cable Damage Detection' dataset from RoboFlow show that the improved YOLOv8 model achieves high accuracy, recall, and precision for detecting 'thunderbolt' and 'break' types of cable damage. Compared to the original YOLOv8 model, the improved version has significantly better performance, highlighting its practical value and strong generalization ability. The paper also discusses the strengths and weaknesses of the proposed method and suggests future research directions.This paper presents an enhanced version of YOLOv8 specifically designed for detecting damage in overhead power lines. The enhancements include an adaptive threshold mechanism, a novel convolution method called GSCov, and a lightweight network structure called Slim Neck. The adaptive threshold mechanism dynamically adjusts the detection threshold based on the brightness and contrast of the input image, improving the model's robustness. GSCov balances the model's running speed and accuracy by introducing a flexible way of information interaction through grouping and random permutation. Slim Neck reduces the model's complexity and computational load while maintaining good performance. Experimental results on the 'Cable Damage Detection' dataset from RoboFlow show that the improved YOLOv8 model achieves high accuracy, recall, and precision for detecting 'thunderbolt' and 'break' types of cable damage. Compared to the original YOLOv8 model, the improved version has significantly better performance, highlighting its practical value and strong generalization ability. The paper also discusses the strengths and weaknesses of the proposed method and suggests future research directions.