I was wondering if there is any way that I could split a big image into smaller ones by laying a grid and automatically splitting and saving individual grid units?
At this point, no, that ability does not exist. However, depending on the grid size, you could create a mask of each individual grid element within the image, identify it as an object and then crop your desired image into that object (removing the blanks rows and columns). Thereafter, you can perform any image manipulation/identification you want as you would with any other image.
We have added this to our long-term To Do list and will post here if the feature is added.