Manual toll collection systems are obsolete due to time, fuel, and pollution issues and need to be replaced by new and better alternatives. Traditionally, governments have always employed people to collect toll, but the manual labor isn't much eff ective when it comes to monitoring and effi ciency. We took this problem and researched out an eff ective solution i.e., "Electronic Toll Collector Framework" which is a framework mainly for collection and monitoring of the toll fees collected by the toll plazas in the vicinity of metropolitan cities like Lahore or Karachi. The software can generate toll tax based on vehicle type. Additionally, it can also generate daily/monthly/yearly revenue reports. The framework can serve other purposes like monitoring of vehicles (by the law enforcement agencies) and generation of analytics. It can also serve as a backbone for the government departments who are having a hard time monitoring the revenue generated by the employers. There are two operational modes of the framework (partly manual and automatic). The partly manual approach uses TensorFlow backend, and the automatic approach uses Yolov2 backend. This work will be helpful in guiding future research and practical work in this domain.