Vehicle Types and Colors Detection for Amber and Silver Alert Emergencies using Machine Learning




Kesava Pillai, Uma Kamatchi

Journal Title

Journal ISSN

Volume Title



The National Center for Missing & Exploited Children estimated that 145 AMBER Alerts were issued in the U.S. involving 180 children in 2019, where 85% had involved vehicles, and in Florida, 136 Silver Alert was issued in (2008-2009). The details of broadcasting in Amber and Silver alerts are color, type of the vehicle, vehicle license plate numbers, and car brands. Vehicle types include six classes as Truck, Bus, van, SUV, Sedan, and Motorcycle. Vehicle colors include eight classes: green, blue, black, white, gray, yellow, cyan, and red. The system works once the Amber alerts are issued; the time taken to recovery is more than twelve hours for children in seventeen cases. More recovery time is due to the time is taken for someone to recognize the vehicle involved in child abduction and authorities to bring back the child or elderly safe back home. This thesis research aims to design a Deep Learning model to classify vehicle colors and types faster and accurately. First, Shallow CNN and Deep CNN, inspired by VGG16, are designed for vehicle colors and classification types. Second, implementation of the pre-trained YOLO object detector to detect vehicles in images. Extraction of license plate characters and numbers using image processing techniques with OpenCV and Python will help identify possible License Plate matches. The whole process will help extract colors and types of vehicles from the child abduction involved, bringing child and elderly person safe home.



Vehicle classification, Machine learning, Detection, License plate, CNN, Amber alert, Silver alert


Kesava Pillai, U. K. (2021). <i>Vehicle types and colors detection for Amber and Silver alert emergencies using machine learning</i> (Unpublished thesis). Texas State University, San Marcos, Texas.


Rights Holder

Rights License

Rights URI