| Abstract | Nowaday, to make a system to process a bunch of video from cameras, we need to have a Video Management Software (VMS), which require a big structure, and also big bandwidth because VMS itself need to refresh every single image (frame) of video. To store images in the storage, and be a center of all video viewer client, video analytic server. So, after VMS get all the images from all of the camera, the VMS need to send those images to all client, and analytic server. If we have a lot of cameras, this require such a big bandwidth. Not only bandwidth and structure are the issues, but power consumption is also the issue. In conventional way, every single device is plugged to the electricity from the factory. Suppose we have a million of the camera all over the town, the amount of electricity usage would be so high. Configuration on those VMSs are also hard. High learning curve is needed to understand how to config them.
To have more efficiency technique, we need to reduce whatever we do not need. Such as VMS, and bandwidth usage. So first, instead of record every single image from each camera, we do record only the set of images we are interested instead. But to do that, the part of the video processing would be on the camera, or somewhere close to the camera instead of the server. This require the system to have a microprocessor and application to process the image from the camera. This would make more efficient way for the bandwidth usage. And using solar cell with the battery to power the camera, store the energy in the battery for the time solar cell cannot generate the power for the camera.
|