While moving the background image instead of the player is a good idea, you may need to find a better method (for example, what happens when you have enemies moving round the level too?)

The method I would use would be to store the player's 'h' and 'v' position in variables (properties), and when the player moves, adjust these h and v variables, and not the sprite's h and v position.

Then, you can calculate where the player's sprite should be displayed, by adding his 'h' and 'v' variables, to the top-left position of your background image.

This way, you can then move your background image so that the player stays roughly in the middle of the screen, but in your programming, your code is 'theorectically' moving the player instead of the background. This will make it much easier for you to move on to your current problem which is the collision detection.

There are a lot of ways to approach this (collision detection), it's really a very large area and probably can't be answered as a single question. Instead I would advise that you do some google searches for phrases like:

"2d collision detection"
"tile based collision detection"
"lingo collision detection"

And read around the subject that you are tackling. You'll find there are many different methods available, and you'll have to decide which method suits the kind of game you are making.

hope this helps!

- Ben