This paper presents techniques and methodologies for object recognition and cooperative pick-and-place of a single spherical wheeled mobile robot (SSWMR) with seven-degrees-of-freedom (7-DoF) dual arms. Both arms are cooperatively used to achieve a pick-and-place task by solving the inverse kinematics of both arms and performing on-line motion planning in real time. To maintain the SSWMR in a stable state to carry out the desired task, the robot is equipped with a lifting platform using four legs to support static stability. Image processing and pattern recognition methods are used to develop application programs based on the OpenCV and OpenNI library to recognize and locate desired objects to be grasped. Simulation and experimental results are used for successful illustration of the proposed techniques.