SCIENTIFIC ABSTRACT: We plan to make broad-band and CVF maps of supernova remnants in order: to study the origin of the dust responsible for the IR emission seen by IRAS in young SNR (circumstellar, interstellar or SN condensate?), to study the composition of this dust (importance of Very Small Grains), to study the heating mechanism of this dust (for example collisional heating for the VSG), and more generally to tackle the problem of grain formation and destruction in SNR (also in SN 1993J). We also plan to study the interaction of a SN wih a molecular cloud (RWC 103 or IC 443). OBSERVATION SUMMARY: We plan to map completely 4 Young Supernova Remnants: CAS-A, Kepler, Tycho and Crab, with the ISOCAM camera. Priority 1 will be given to CAS-A; the total IRAS flux at 12 micron is 17 Jy and the peak flux is of 0.6 mJy/arcsec2. Four frames with a PFOV of 6" are needed to map the entire remnant (plus 2 frames for the background) (raster mode 3,2; steps of 28 pixels). The LW8 filter will be used first with a 4.5 minutes exposure time. A signal to noise ratio as high as 150 should be obtained at the peak. The total time for the observation is 27 minutes. On the brightest frame of the remnant, we plan to use in addition to the LW8 filter, the LW9 filter and 2 CVF positions (10.5, 11.3 microns). With 9 min on each filter, a S/N ratio of the order of 100 is expected; total time: 27 min. Given 3 min for pointing and given slew and concatenation time, we arrive to a total of 1 hour for this set of observations. In priority 3, we would also like to observe the brightest frame with the NeII line CVF position (12.8 microns) and the reference positions at 12 microns and 9.8 microns. With 13 min per filter position, we end up with a total time of 45 min. Exactly the same observing strategy, will be used for Kepler. Kepler is 5 times fainter than CAS A; still, we should obtain a good S/N on the brightest part of the remnant. Total time spent: 1.45 hour. Tycho is 5 times fainter than Kepler and much larger, so that only one filter (LW8) will be used. The complete mapping (15 frames; raster 5,3; steps: 30 pixels) with 5 minutes per frame will require 75 min, and the brightest parts will be observed; for fainter parts, we plan to degrade the spatial resolution to gain in terms of S/N . The Crab nebula (spring launch) is a completely different kind of remnant: a plerion, in which synchrotron emission may be important. We plan to map it in the LW2 and LW9 filters with 4 minutes per frame and per filter to obtain a S/N around 100 at the emission peak. Given the number of frames needed for the complete mapping (15 frames; raster 5,3; steps 30 pixels), 120 min are required. RCW 103 (autumn launch or IC443 spring launch) is interesting to test the interaction of a SN with a molecular cloud. We plan to use the Ne II tracer at 12.8 micron (CVF point) over 15 frames (raster 5,3; steps 30 pixels) to cover the whole remnant, with 5 min per frame. Total time 75 min (only 60 min if IC443). Note that to make the mappings, we plan to use the raster scan mode with scanning along the axis of the telescope. For SN 1993J, we plan to use 5 filters (LW1, LW2, LW7, LW8, LW9) with 5 minutes exposure per filter; (the time being driven by the detector transients). Total time with overheads: 30 min. CONCATENATION: The observation on CAS-A (priority 1) have to be concatenated, as all the observations are needed to interpret correctly the data. Furthermore, there is a gain in overheads.