This can be done by writing code that will check for the NaN values in the image matrix and set it black. For example, create a 5X5 matrix 'a' containing some NaN values and some '0' values. Get minimum and maximum values of that matrix and also location of all 'NaN' values. Then replace all NaN with values which are less than minimum number present in that matrix. Then redraw the colormap and plot the image.
a =[
0 0.8843 NaN 0.3900 NaN
NaN 0.5880 0.8256 0.1117 0.4950
0.8620 0.1548 NaN 0.1363 0.7476
NaN 0.8999 0.3185 0.6787 NaN
0.5144 0 0.5341 0.4952 0.8507
];
minv = min(min(a));
maxv = max(max(a));
a(isnan(a)) = minv-((maxv-minv)/5);
ddd=[0 0 0;jet(10)];
colormap(ddd);
imagesc(a)
