(Yicai Global) Sept. 8 -- A team of researchers at Peking University's School of Electronics Engineering and Computer Science is developing shape recognition technology for vehicles.
The technology will be able to accurately detect, record and analyze vehicles' external features, automatically determine their model and retrieve their registration records, Chinese media outlet Feng reported yesterday.
Dubbed Repression Network (RepNet), the equipment will record two types of features: general features, such as color, brand and model; and visible defects and damage. The system will have a camera designed to capture high-definition pictures of cars, and process and identify images after receiving the vehicle's information.
It's still too early to tell when RepNet can actually be used, as some people may consider it an invasion of privacy, one researcher said. One thing is for sure though, if the police can use the system to screen images of vehicles recorded by video surveillance systems, it would go a long way toward helping car theft investigations. If a massive image and video database can be created based on the country's extensive public security camera network, it would make car recognition much easier.
License plates hold important information about cars, however, numbers, letters and characters on plates still cannot be reliably recognized by existing plate recognition systems, which is why the team set about developing RepNet to identify vehicles based on their appearance.