Deep Learning Based Vehicle Make, Model and Color Recognition Using License Plate Recognition Camera Images

In this study, we propose a deep learning based vehicle make, model and color recognition method using license plate recognition cameras installed in highways. In the proposed approach, vehicle localization is carried out using single shot multi-box detector (SSD) model. Next, we utilize a deep learning based make, model and color classification model on the vehicle region. Data sets of different shade, light, reflection and other lighting effects have been created to be used in the training and testing stages of the proposed methods. Vehicles make and model classification model was tested using 6238 NIR and RGB images, while the color recognition model was tested using 3588 RGB images. Test results of vehicle make, model and color classification methods were compared with the performances of the existing cameras on roadways today. Proposed model achieved a 94% accuracy rate in make & model recognition and 92 % accuracy rate in color recognition tasks.