Batch conversion of fits files to .png images, dynamic range and histogram equalisation

My sample consists of almost a thousand galaxies. It's that inconvenient sample size where you would like to inspect each result visually, but it already sucks. First of all, I wanted to see if the inpainting code makes sense and leaves no strange artifacts. To avoid opening each .fits file manually with ds9, I scripted some quick loop to save all the fits images as .png files. It uses scipy.imsave():

scipy.misc.imsave('img/output/filled/'+ID+'_imsave.jpg', outputImage)

However, I wanted to stretch the pixel intensity values and increase the dynamic range of the output images. This code was useful. It tries to make all intensities equally common and hence increases the contrast.