Robotic systems for unstructured environments need to cope with rapid changes in time and space, inherent uncertainty, and generate unpredictable and dynamic situations, resulting in lack of information, due to inadequate sensor accuracy and computational limitations. Introduction of a Human-Operator (HO) into the system can help simplify the robotic system and ensure improved performance. In the current study we developed an objective function for different levels of HO-robot collaboration in a target recognition task. The objective function considers several HO and robot parameters, which contribute to the overall performance in each level of collaboration. Results indicate that HO-robot collaboration improves system performance in many conditions. The optimal collaboration level depends on image complexity and on robot and HO characteristics. Some of the latter have to be assessed through behavioral research on HOs.