work with 8-bit image in retile.py and tile_utils.py

[tnwillia]

I've successfully labeled both panchromatic satellite imagery and UAS thermal imagery. However, when I try to use these image to create a library, the panchromatic images (8-bit depth) fail. The UAS imagery (24-bit depth) completes the process. An example of the error is shown below. Is there a way to adapt this to grayscale imagery? Thanks.

(dl_tools) C:\WORK\machineLearning\dl_tools>python create_library\retile.py -t 96 -a 0.9 -b 0.1
Searching for labels in all 9 files
Working on C:/WORK/machineLearning/dl_tools/allLabeledImages/IN_pan_2010apr141_mres.mat
Generating tiles from dense class map ....
Traceback (most recent call last):
File "create_library\retile.py", line 143, in
Z,ind = sliding_window(imread(fim), (tile,tile,3), (int(tile/2), int(tile/2),3))
File "C:\WORK\machineLearning\dl_tools\create_library\tile_utils.py", line 56, in sliding_window
'a.shape, ws and ss must all have the same length. They were %s' % str(ls))
ValueError: a.shape, ws and ss must all have the same length. They were [2, 3, 3]