When using a picture to composite into like these images you need to match the lighting almost exactly to hide the cg element in the scene. The main things I look for in the photo are the direction of the shadows, softness of the shadows, the light intensity, and the color of light. In the compositions I used two or three area lights to get the same effect as using radiosity, which helped with render times. I set up a warm light for the main light, and then I used cool lights for fill in.

In the case where I need the cg element and real photo to interweave, I use polygon blockers to block out the cg element in that area. To do this I take the same photo I am compositing into and create a polygon in modeler based on the shape of area I want to block. Then I would front project the image onto those polygons in layout to get parts of the image to be overtop of the cg elements. You can see an example of this in my third image with the box car in the meadow. The bushes in the front of the box car are the blockers with front projection maps on them. After rendering the image, I don’t consider it cheating in any way to go into Photoshop and make the colors more vibrant or dull depending on the look you are going for. It can only help your image.