We developed a new computerized CCD video light microscopy system for obtaining three-dimensional (3-D) RGB measurements of shape (with color texture) and growth of intact plants under various conditions and over a wide magnification range. We used a modified shape-from-focus algorithm that incorporated a linear regression operator for 3-D reconstruction of images of petunia seedlings, which have a coarse texture. This algorithm was unsuitable for processing images of specimens with glossy texture (e.g., plant cells), but this drawback was mitigated by illuminating the cells in a checked pattern. The algorithm also reduced incorrect range estimates, which are due to defocused areas around the object's edge in original images. Within only a few minutes, the presented system can yield a series of original color images and reconstruct 3-D images incorporating color texture. All operations of our new system are carried out automatically, and the results are displayed by using in-house developed software that allows observation of the reconstructed 3-D image from any direction on the upper side. At angles from 10 to 80° to the objective's face, the surface area can be estimated to within an error of 5%.