Abstract :The classification of sound has become an important research topic in recent years. Although the classification of the sound is difficult due to the nature of the sound, the development of computer hardware and the high performances of deep learning have made it possible to classify the sound. Classifying sound is an important area of research, and classification of environmental sound data can be beneficial in many areas. Classification of environmental sounds can be used in areas such as multimedia, security perimeter control, and sound surveillance. In daily life, many different methods are used to process sound data and make automatic inferences from this information. Among them, deep learning-based approaches are systems that have high performance and accuracy in processing large audio data, give fast results and have the potential to be used. In the study, the UrbanSound8K dataset was trained on magnitude mel-spectrograms extracted from the audio. Classification was carried out using supervised learning techniques using 10 different environmental sound data classes. Keywords : Deep Learning, Classification, Audio Processing, CNN
