Depth-Based Bin Pose Detection for Robotic Picking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robotic bin picking systems face inefficiencies in estimating the pose of bins due to technical challenges such as dynamic bin positions, noisy sensor data, and lack of accurate system information, which can impact grasp computations and overall performance.
Innovation Solution
An autonomous system equipped with a depth camera and processor that generates a segmentation mask of the bin to identify its pose by fitting mathematical models to the bin's boundary points, allowing for precise bin pose estimation without the need for artificial markers or detailed CAD models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current approaches are used to estimate bin pose, then the system can operate with simple sensors, but the pose estimation accuracy deteriorates due to dynamic bin positions and noisy sensor data
Solution Approach 1:
The patent segments the bin detection problem into multiple processing stages: depth image acquisition, segmentation mask generation, boundary point identification, and model fitting. This segmentation allows each stage to be optimized independently, improving overall pose estimation accuracy while maintaining system reliability in dynamic environments.
Solution Approach 2:
The system performs preliminary actions by generating a segmentation mask and identifying boundary points before final pose estimation. This preliminary processing filters out noisy data and focuses computation on relevant features, enhancing both accuracy and reliability when dealing with dynamic bin positions.
2Measurement precision
If detailed CAD models or artificial markers are used to improve bin pose estimation, then measurement precision improves, but device complexity and ease of manufacture worsen
Solution Approach 1:
The bin itself serves as the reference object without requiring external markers or detailed CAD models. The system extracts geometric features directly from the bin's visible boundary in the depth image, allowing the bin to provide its own pose information through its natural geometry. This self-service approach maintains high measurement precision while avoiding the complexity of marker systems or comprehensive CAD modeling.
Solution Approach 2:
Instead of using detailed CAD models, the system creates a simplified geometric representation (copy) of the bin based on observed boundary points. This copied geometric model is sufficient for pose estimation purposes and can be generated rapidly from sensor data, reducing system complexity while maintaining adequate precision.
3Productivity
If traditional bin picking approaches are used, then the system is easier to implement, but productivity deteriorates due to inefficient pose estimation
Solution Approach 1:
The system performs partial action by focusing pose estimation only on the necessary geometric features (boundary points of the bin) rather than complete 3D reconstruction. This selective approach reduces computation time significantly while providing sufficient information for efficient bin picking operations, thereby improving productivity without excessive processing overhead.
Data Source
AI summary
It is recognized It is recognized herein that current approaches to robotic picking lack efficiency and capabilities. In particular, current approaches often do not properly or efficiently estimate the pose of bins, due to various technical challenges in doing so, which can impact grasp computations and overall performance of a given robot. The pose of the bin can be determined or estimated based on depth images. Such bin pose estimation can be performed during runtime of a given robot, such that grasping can be enhanced due to the bin pose estimations.