<p>With the widespread use of unmanned aerial vehicles (UAV), their safety issues have become increasingly prominent in recent years. Therefore, UAV detection and identification technology has become a hot spot. Radar-based methods make it challenging to monitor low-flying UAVs, and video-based methods require high imaging quality. The acoustic signal-based UAV detection method can compensate for the shortcomings of these traditional UAV detection methods. This paper proposes an integrated learning model based on multi-scale convolution and global local attention by processing audio signals. The model can perform accurate UAV identification through UAV audio signals and aims to complement the shortcomings of other methods. The model proposed in this paper adopts an integrated learning framework, which can directly process the raw audio signals of UAVs without manual feature extraction. The proposed model consists of two first-level expert models and a meta-classifier. Firstly, the two first-level expert models perform feature extraction on the data separately. Then, the obtained classification results are inputted to the meta-classifier. Then, the meta-classifier integrates and fuses the results of the first-level models and finally outputs the results of UAV monitoring and recognition. The two first-level expert models add a multi-scale global local attention module based on the residual and depth-separable convolutional structures. The method in this paper is compared with other methods for processing one-dimensional signals on a self-created UAV dataset. Experiments verify the effectiveness and superiority of the model.</p>