I am very new to programming so I wanted to ask how difficult this program I would be interested in making would be. I know this program is already way out of my league as I am a novice but it would be interesting to know how difficult such a program would be, especially if it turns out to be rather simple.
I want to make a program that reads jpegs or any sort of image of satellite images & can identify a manmade object (building or structure).
The 'algorithms' I thought of using are; 1st to scan over the topographic map of the image & compare the contour lines to identify areas that have a contour difference of no more than 10 metres over an area of 200 square metres, ie the average size of a house/building. This would, through deduction, eliminate mountain ranges & oceans etc.
On the remaining areas of the satellite image that have been identified as suitable to contain a building or structure, the program then scans for right angles ( a 'shape' not commonly found in nature, but usually always in manmade objects such as the buidlings themselves but also the shadows of the buildings)
I thought why shouldn't the program just scan the whole image for right angles & forget about the contour comparision function; but I thought that the program is probably going to return alot of false positives so it wouldn't hurt to have another piece of criteria in there. I would also put it in for the speed of the program, meaning it will take less time if the program has to search less area.
Is this a rather complex program to write in your opinion? Also how could i write the program to search for right angled shapes, how would I do it if, say, I was using C++?