Automatic Vehicle Number Plate Detection and Recognition (AVNPDR) system can be used to control and implement traffic rules effectively. A number of machine learning classifiers are employed for numerous object detection and recognition purposes. In this work, three classifiers viz., Haar cascade, Local Binary Pattern and YOLO v5 are used to detect and localize the number plate present in an image. The performance of these classifiers and their suitability for AVNPDR is studied for real-time implementation. In this work, the machine learning models are trained with number plates whose characters are present in a single line. After localization of the plate, the characters present in the number plate are segmented using image processing techniques. Finally, the characters are fed to an optical character recognition system PyTesseract, which recognises the characters and displays the recognised character as a string. The imple-mentation of the three models in real-time implementation is carried out with the help of Raspberry Pi Model 3B and their performances are compared. Among the three classifiers, Haar classifier ranks the best on accuracy (99%), precision (100%) and F1 score (99.3%) followed by YOLO and LBP.