Increasing number of vehicles on the streets has made it very difficult to track every vehicle, which has resulted into exponential rise in violation of traffic rules. It is very difficult to manually keep track of all the vehicles and keep them in check. Due to human limitations, intelligent and smart monitoring has become a topic of discussion. For every vehicle to be acknowledged, number-plate detection is very necessary. However, many vehicles in a single image and varying lighting conditions present challenges to automatic number-plat recognition systems. Varying shapes and sizes of characters on number-plate presents additional difficulties in Indian context. To address these issues, this work proposes to employ inception version 3 (v3) model for classification of number plates form input images. The study then proposes to employ single shot detection (SSD), which is one of the best available architecture to detect multiple objects at once, resulting in faster and more accurate detection. At a later stage, each character in the number plate is recognized by Tesseract optical character recognition (OCR) engine. The experimental results presented in the study validates the effectiveness of the proposed work.