Camera Possibilities
Mick Walters (UH)
I have been looking at the spec for the VC-C50i camera and it seems that it is backwards compatible with the vcc4 (which is probably why the Piopneer robot people chose it?). The input/output connections are:
Installation of your PTZ camera is easy because of the connection options the VC-C50i provides. The camera's secure and sturdy Multiple Connector Interface and Connector Block give you S-Video out, RS-232C in and out, as well as connections for Alarm out, Sensor in, and External Light Trigger. Plus, for those seeking built-in video and control connectors on the unit, the VC-EX3 connects to the VCC50i providing direct access to S-Video, RS-232 in and RS-232 out. Your VC-C50i also comes with BNC video output and a supplied BNC/RCA video adaptor for even greater flexibility. With all this flexibility, you can get it connected -- and get it working for you right away.
Incidently, there are two main variants of the Cannon PTZ cams. the VC-nn series which are primarily analogue output devices (no network connection), and the VB-nn series which are primarily network cams, but some include an aux analogue video output.
We have also had some experience at using network video cams. A network camera is quite convenient for setting up remote cameras where there is a wired or fixed network in place, but from my (limited) experience with streaming video over a wireless network the response seems to be quite slow (a delay of .5sec is usual?) and the bandwidth limited (so no high frame rates or high res). This may not be a factor with fixed computer based network/vision applications, but for mobile robots it can cause problems. I am not sure, but I suspect the network camera software drivers are proprietory or specific to the network camera being used, so you need to make sure that if you use Linux, the camera is supported.
I have been looking at the IEEE 1394 specification in more detail and I think it has several advantages for us in terms of connecting camera hardware. First it is an international standard, and has a unified set of drivers for windows/linux etc. which operate and download digital video from a range of different cameras. For an overview of the Interface see:
http://www.theimagingsource.com/en/resources/glossary/entry/ieee1394.htm
Incidently, this manufacturer supplies a range of IEEE 1394 cameras and more importantly an IEE1394 interface which converts any source of analogue video to IEEE1394 compatible dv. See for cameras:
http://www.theimagingsource.com/en/products/cameras/firewire_color/
and convertors:
http://www.theimagingsource.com/en/products/converters/dfg13941e/overview/
In terms of connectivity, we could in principal connect any video source to a IEEE 1394 interface. The range of cameras would emcompass dedicated IEEE1394 cams, camcorders with firewire/ilink (just other names for IEE1394 interface)and via an IEEE1394 converter any other analogue video source such as the old VCC4 and newer VC-C50i variant used on the Peoplebots. Both cameras use the RS232 port only to control the PTZ mechanism, which would still be operated through the ARIA system as before. The analogue video could still be captured by the PeopleBots on-board capture card, so even the PeopleBots vision software could still be used. However by using an IEEE convertor to interface to any external computer, this provides several advantages. The IEEE1394 cables can be long (we have used 5m to feed signals to a computer). The interface is standard across windows and linux platfoprms, so software written for one IEEE1394 camera should work almost unmodified with any other IEEE1394 camera (apart from allowing for different scanning resolutions).
To summarise, I think that the actual camera different partners use is not so important. The only important thing is that it can be connected (either natively or via a convertor) to a standard IEE1394 computer interface for streaming the digital video for further processing.
What do people think?




