In this letter, we propose a new method to fully control complete 4-image-DoF manipulation of laparoscopic instruments [with remote center of motion (RCM) mechanism] based on the geometric features of a designed marker in a 2-D image. Our marker encodes the configuration of the instruments by computing geometric features among the projected image points from segmented areas in hue-saturation-value (HSV) space. We can then construct an image geometric feature vector to locally characterize the configuration of a laparoscopic instrument. Furthermore, we design an image-based kinematic controller to asymptotically track a planned trajectory using the constructed feature vector as the feedback. We evaluate our integration of rotation distinguishing marker and kinematic controller by several experiments in terms of illumination-invariance, rotation angle accuracy, and controller performance.