In response to the challenge of identifying small targets in complex environments,this paper proposes an airborne radar target recognition method based on the spatial and temporal distribution characteristics of power towers. The algorithm constructs a fusion model using convolutional neural networks(CNN)and long short-term memory (LSTM)networks,and introduces strategies such as sample augmentation and weighted entropy to effectively mitigate the shortcomings of models due to imbalanced samples.Validation is conducted using real flight data,demonstrating that the proposed method reduces false alarm rates to 20.5%and inereases accuracy to 80.15%in complex backgrounds, offering higher precision and robustness compared to traditional radar target detection methods and typical CNN/LSTM networks.Considering practical conditions of helicopter low-altitude flights,the proposed method leverages the agility and adaptability of artificial intelligence technology to achieve high-precision recognition of power tower line targets by airborne collision avoidance radars.