So you need to do things step by step
- read the image
- crop (and +repage) the part of the image wanted
- resize to the new size in pixels
- set a geometry position
- read destination image
- swap two image so the croped image is second (overlay) image
- set a geometry position
- compose the two together
- save the result
The steps can be modified for non-rectangular areas (masking instead of cropping, or for warping and distorting the image into a new shape or location before overlaying, but essentually the same basic steps should be followed.
You can break up the steps into separate commands, or insert other modifications, but the basic steps remain.
NOTE : we are not here to do the work for you, bt help and guide you to the best solution. With actual examples we can generate actual commands, but it is better if you actually understand what individual steps should be taken.