Visual surveillance is increasingly prevalent today but the privacy issues of individuals involved in surveillance videos have not been dealt with adequately so far. In most cases, if not all, the chief purpose of placing a camera can be served without knowing the identity of the individuals involved unless the activity is of some predetermined kind. One needs to transform these videos to protect identity of individuals involved possibly at the source camera itself. In this paper we present the De-Identification Camera, which is a scalable, low cost and real-time solution to the privacy protection issues. Our main contribution lies in proposing a privacy protection architecture which transforms the video at the camera level itself, We also present and implement a de-identification pipeline which is suitable for real-time implementation. We implemented the system on a Texas Instrument OMAP4 based embedded platform and was able to de-identify videos in real time, transforming the video on the camera itself ensures protection from various attacks. We also address issues like data utility of surveillance videos by making this solution customizable. We propose that such systems can replace the traditional surveillance cameras in the future by providing all the surveillance and privacy protection solutions on hardware, probably with few performance upgrades.