A color stereo vision system consisting of two Firewire (1394) cameras and computer vision analysis software running on a PC is used to reconstruct digital elevation models (DEM's) from soil surface stereo images. A laser grid pattern generator is used to project points onto the soil surface to assist in defining reference points between the images in areas where correspondence between images is difficult to determine. The OpenCV image processing library and custom software written in C++ are used to calibrate the cameras, perform image rectification, generate disparity images and reproject pixels to three-dimensional coordinates. System performance is compared with an existing single camera line laser scanner for measuring topography in a laboratory meter-sized soil box. DEM's are generated with and without using the laser grid pattern in the image processing to access the laser grid generators effect on DEM accuracy. Results from a field test on a hillslope scale will be discussed.