Unmanned aerial vehicles (UAV), or drones, offer immense potential but also pose major security concerns due to their accessibility and misuse. Hence, effective detection and identification of drones are crucial for mitigating these risks. This study explores the application of different preprocessing techniques and deep learning models for the identification of drones and the detection of unknown drones by their acoustic signature. This work combines various acoustic files from numerous sources and utilizes data augmentation techniques as needed to train the deep learning models adequately. This study is focused on Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Convolutional Recurrent Neural Networks (CRNN), analyzing the performance of each with Mel spectrograms. Our findings demonstrate the efficacy of deep learning models in achieving promising results for the audio identification of drones. Notably, Mel spectrograms consistently yielded the best overall classification results for all three models, CNN, RNN, and CRNN. Additionally, our application of entropy thresholding proved successful in detecting unknown drones, resulting in robust Mel models that were able to classify both known and unknown drones fairly well.