MUMBAI, Nov. 22 -- The state government is planning a major technology upgrade of the city's CCTV infrastructure, introducing artificial intelligence (AI)-enabled cameras capable of facial recognition, thermal imaging and detecting traffic violations, including two-wheeler riders jumping signals. The government has appointed Hyderabad-based GVPR Engineers Ltd to implement the third phase of the Mumbai City Surveillance Project (MCSP), at an estimated cost of Rs.2,098.92 crore. The upgraded network is expected to go live in February 2027 and will also allow the police to access CCTV feeds from private establishments such as five-star hotels, corporate offices, malls and large housing complexes. The new system will mark a shift from conventional surveillance, which largely relies on monitoring footage and retrieving recordings after an incident, towards automated detection of suspicious activity and violations. Among the planned features are facial recognition, automatic number plate recognition (ANPR), thermal imaging and red-light violation detection. An important addition will be cameras capable of identifying two-wheelers jumping traffic signals and detecting helmet violations. At present, the city's traffic enforcement system primarily captures number plates of four-wheelers for issuing electronic challans. "The third phase will see surveillance network upgradation with advanced, top-end cameras and AI-enabled features such as facial recognition, thermal imaging, automatic number plate recognition and red-light violation detection. It will also introduce detection of two-wheelers jumping signals, which was missing until now," a home department official said. The Mumbai Police's command and control centre will be connected to feeds from participating hotels, malls, corporate establishments for tracking suspects or investigating crimes....