Pose estimation of objects using digital image processing for pick-and-place applications of robotic arms

Abstract EN

Robot Vision is one of the most important applications in Image processing.

Visual interaction withtheenvironment is a much better way for the robot to gather information and react more intelligently to the variations of the parameters in that environment.

A common example of an application that depends on robot vision is that of Pick-And-Place objects bya robotic arm.

This work presents a method for identifying an object in a scene and determines its orientation.

The method presented enables the robot to choose the best-suited pair of points on the object at which the two-finger gripper can successfully pick the object.

The scene is taken by a camera attached to the arm's end effector which gives 2D images for analysis.

The edge detection operation was used to extract a 2D edge image for all the objects in the scene to reduce the time needed for processing.

The methods proposed showed accurate object identification which enabled the robotic to successfully identify and pick an object of interest in the the variations of the parameters in that environment.

A common example of an application that depends on robot vision is that of Pick-And-Place objects bya robotic arm.

This work presents a method for identifying an object in a scene and determines its orientation.

The method presented enables the robot to choose the best-suited pair of points on the object at which the two-finger gripper can successfully pick the object.

The scene is taken by a camera attached to the arm's end effector which gives 2D images for analysis.

The edge detection operation was used to extract a 2D edge image for all the objects in the scene to reduce the time needed for processing.

The methods proposed showed accurate object identification which enabled the robotic to successfully identify and pick an object of interest in the scene.