Bin Picking

Bin picking is a robotic process where a vision system and robotic arm work together to identify, locate, and pick objects from a bin or container. The system uses advanced algorithms and sensors to detect the shape, position, and orientation of the objects, even in cluttered or unstructured environments. Bin picking is commonly used in industrial automation to improve efficiency and precision in tasks such as sorting, assembly, and material handling.

CHALLENGES

Incorrect Pose Estimation

Incorrect pose estimation in bin picking occurs when the vision system inaccurately detects the position, orientation, or alignment of an object within the bin. This can lead to the robot either missing the object, failing to grasp it correctly, or even causing collisions with other objects. Such errors reduce the efficiency of the process and can lead to damage or production delays, making accurate pose estimation crucial for reliable and precise bin picking.

Failed to Pick Parts

Failing to pick parts in bin picking occurs when the robotic system is unable to successfully grasp or lift an object from the bin. This failure can result from various factors, such as poor pose estimation, incorrect gripper alignment, part collision, or an unsuitable grasping point. When the robot fails to pick a part, it disrupts the automation process, leading to inefficiencies and potential downtime in production.

Failed to Place Parts

Failing to place parts in bin picking happens when the robot is unable to correctly position or release an object at the intended location. This failure can be due to inaccurate pose estimation, misalignment of the gripper, or issues during the release mechanism. As a result, the part may be dropped, misplaced, or even damaged, disrupting the workflow and decreasing overall efficiency in automated tasks like sorting, assembly, or material handling.

POSESTREAM SOLUTIONS

Robust Pose Estimation

PoseStream offers advanced solutions for correct pose estimation by leveraging deep learning and computer vision technology to accurately detect the position and orientation of objects in real-time. PoseStream is designed to handle challenging environments, such as cluttered or partially occluded scenes, ensuring precise object localization. By providing accurate pose data, PoseStream enables robotic systems to reliably pick, place, and manipulate objects with high precision, reducing errors and improving efficiency in bin picking and assembly tasks.

Efficient Part Picking

PoseStream does grasp planning in real-time, finding all available grasps of located parts while ensuring that the robot gripper will not collide with surroundings. The system calculates the optimal grasping point for each object, even in complex environments with clutter or occlusion. This precision enables robots to securely and effectively pick parts until the bin has been emptied.

Precision Assembly / placement

PoseStream enabes in-hand positioning of parts while the robot is moving. After locating the object with pose stream, the robot tool center point can be adjusted to the actual part position. This enables precision assembly with up to 0.3 mm accuracy.