Automatic aircraft recognition using convolutional neural Network

The paper presents the possibility of utilisation convolutional neural network for aircraft classification by their type. Main purpose of the study was to use a common deep learning network and modify it to correctly classify three types of general aviation aircraft. Differentiation is made based on their low quality picture with black outline on white background. Database utilized in this study is small compared to common CNN databases and results ought to be taken only as a trend. Research consisted of choosing right parameters of network to make the recognition as accurate as possible. 20 samples have been produced to evaluate accuracy of the software and eliminate deviations. Conclusions and issues have also been described.