Service robotics must balance the design and development of functional units with simple human-robot communication to generalize its use and service to the public. This paper presents an implementation that combines color and shape detection to recognize four figures with specific shapes and colors. Each figure represents a basic control instruction that will be sent to a differential drive robot. The algorithm uses images extracted from a webcam using MATLAB computer vision libraries, and then processed by methods such as binary mask generation, edge detection with a Canny filter, edge dilation, and detection of straight lines with the Standard Hough Transform. At the same time, the instructions are executed in the robot model with the help of packages developed with ROS approaching real-time control. Tests were carried out in a semi-controlled environment with ambiental disturbances, two different lighting conditions, and digital images to validate our proposal.