Classification of avian radar data by machine learning and developing an educational application

The main purpose of this study is to classify the avian radar data with the artificial neural network (ANN) and to give training of an application to be developed with the obtained classification algorithm. The data used in the study were obtained from an avian radar in Istanbul. Although bird radar is a specialized tool for detecting birds, it can also detect objects called clutter that are not desired to be detected by the radar. In this study, it is aimed to classify the rain which is one of the unwanted echoes by using artificial neural network. For this purpose, training and test data sets were prepared using TPR images for rain-bird classification. The model is written in the Spyder development environment using the python 3.7 language. Systematic experiments were performed to find the optimal number of hidden layers and the number of neurons in each hidden layer. The outcome of the study showed that ANN achieved successful results in classifying radar targets as rain and bird. 85% of the data tagged as rain and 80% of the data tagged as bird are correctly classified.