00100	This is a description of plans for the next few years in vision,
00200	as I see them here.
00300	
00400	cooperation on hardware--sensor cooperation
00500	vision language--cooperation on software
00600	strategies for vision and assembly
00700	visual feedback
00800	planetary exploration
00900	stereo
00950	integrated visual systems
00975	software plans
01000	
01100	manipulation implies some form of depth perception.
01200	strategies imply the need for intermediate level of
01300		language, ie vision language
01400	representation is the key to visual systems
01500		useful representations at all levels are to
01600		be included in the vision language
01700	
01800	
01900	SOLID STATE IMAGE DEVICE INPUT
02000	GE will start delivery in December of 188x244 CID (charge injection device)
02100	solid state image sensors.  These have the following properties of interest
02200	to us:
02300		no blooming
02400		better signal to noise ratio than TV
02500		variable input rate (making easy to match to PDP11 bus)
02600		potential random access
02700		very small
02800		small optics
02900	In implementing such a system, we are negotiating a joint effort with
03000	several other labs to develop a single, plug-in device including
03100		pan,tilt,stereo,color filters, zoom or fovea
03200	The standard device would presumably accomodate several different
03300	basic solid-state sensors, interchangeable.  The largest interface
03400	problem at this point is the mechanical interface.
03500	One possibility for the mechanical system is to use the small
03600	Scheinman arm for the pan/tilt system.  For the first two or three
03700	joints of that arm, the cost would be about $2300 to Scheinman,
03800	without manual controller.  We would need motorized zoom,iris,
03900	focus,filters and stereo.  For those, we can find part of them
04000	available in super-8 movie equipment.  Clearly, the whole of such
04100	a system will cost us about $5000 for the pan/tilt and $3000 for
04200	the zoom,iris, etc.  Having a standard version of such a system
04300	would make it possible for other laboratories to immediately use
04400	these devices.
04500	
04600	These plans bear upon the problem of how to make the PDP11 system
04700	useful for vision.  As we all know, the hardware building for the
04800	PDP11 has gone very slowly, because the memory map, single bit
04900	error correction and memory interface are complex.  Part of the
05000	complexity is the 72 bit wide channel for visual input and display
05100	output.  The bandwidth requirements of the solid state device are
05200	low enough to work on the bus.  If we could get a  low bandwidth
05300	color display device, then there would be no motivation for
05400	extending the memory bandwidth.