Detection of squares, output as svg

Create a python based script that when provided with an image file (and some other inputs - see below) outputs a SVG file with areas of the image outlined. This is to automate the process of making a "key guard" - a piece of perspex that fits on a device (e.g. iPad) to help people with disabilities. An example of final products can be seen here: [login to view URL] An example of SVG file that is expected can be found in the attachments "[login to view URL]" In more detail (pseudo-code): - Read in provided image from command line e.g. [login to view URL] --image [login to view URL] - Detect squares/oblongs from image (e.g. [login to view URL] - 6x4) - Create a SVG path from squares. Curvature of squares can be modified by the user. As well as line thickness & colour. Default: 0.01mm black. - Output final svg to exact size. To do this may require the physical outer size of the screenshot, and DPI of the screen. Maybe not both. (i.e. DPI, WinWidth (mm), WinHeight (mm)) Example: Inputs: Image: [login to view URL] WinWidth: 198mm WinHeight: 148mm DPI: 132 Optional Inputs: Curvature of squares: 4, 5, 6 etc.. (just a value of more curvature) line thickness: 0.01mm (default) line colour: black (default) Output: [login to view URL] NB: This example is purposefully chosen as it may be too difficult to detect the edges of cells in the black areas. If they cannot be detected then that will be a needed compromise. The main body part is the most important. If it can only read in certain image types then that is fine for now too. Other Input Images examples: Accent1 Accent2 Grid1 [login to view URL] Mexpress1 Novachat1 Proloquo1 (These are screenshots taken from a tablet device - this may be an iPad, a windows based tablet or other...)