Without an increase in cropland, agricultural efficiency must be tripled in the next 50 years to sustain the increased demand for food. Controlled environment agriculture (CEA) systems are likely to play an important role in the increase of agricultural efficiency. CEA systems, however, require constant observation because decisions must be quickly made when plants show signs of stress. A visual inspection system that uses a robotic camera system would permit visual access to inaccessible plants in a large hydroponics operation or allows an observer to remotely inspect plants for multiple small or remote CEA operations, whereas a dedicated CEA specialist would be beneficial but impractical under present conditions. This thesis presents a theoretical design for a plant inspection robot. The design parameters, design process, and the system specification necessary to satisfy the design constraints were examined for this system. The design analysis revealed that the major components of the plant inspection robot must be designed sequentially, starting with the imaging system. The imaging system design revealed that the system parameters were governed by illumination, shape and size of the object, and the desired detail. The motion system design was governed by velocity, acceleration, work area, and accuracy. An example design for a system used for visual inspection of 289 romaine lettuce plants was presented. This design was shown to be feasible from the theoretical perspective and could be built from commercially-available components, reducing development time and cost.