Crane-Mounted 3D Object Detection for Large-Area Picking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 3D machine vision solutions are limited in their ability to effectively identify and pick objects in large areas due to restricted spatial range and volume of image data processing, often relying on fixed-position cameras rather than movable 3D image sensors that can cover extensive spaces.
Innovation Solution
The system positions 3D image sensors, such as LIDAR, on a movable crane to generate and process extensive 3D image data, allowing for the recognition of objects across a large area by comparing image data to predefined models and filtering processes to determine accessible objects for picking, with real-time adjustments for optimal object retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If fixed-position cameras are used for 3D machine vision, then device complexity is reduced, but the spatial range and volume of image data processing is limited
Solution Approach 1:
The patent applies the dynamics principle by transitioning from fixed-position cameras to movable 3D image sensors mounted on cranes. The sensors can move through space to capture images from multiple positions, dramatically expanding the spatial range covered while maintaining manageable device complexity through automated positioning systems
Solution Approach 2:
The patent implements another dimension by adding mobile positioning capabilities to the 3D vision system. Instead of relying solely on multiple fixed cameras covering different areas, a single sensor can move through three-dimensional space, adding temporal and positional dimensions to data collection
2Area of stationary object
If movable 3D image sensors are used to cover extensive spaces, then spatial range is improved, but device complexity increases
Solution Approach 1:
The patent applies universality by integrating multiple functions into the movable sensor system. The crane-mounted sensor serves both as a positioning platform and an imaging device, while the system simultaneously performs object detection, spatial mapping, and navigation functions that would otherwise require separate systems
Solution Approach 2:
The system implements self-service through automated sensor positioning and data processing. The movable sensor autonomously navigates to optimal positions, automatically captures and processes images, and identifies objects without requiring manual intervention, reducing operational complexity despite increased device capability
3Measurement precision
If a greater volume of image data is processed, then object identification accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies preliminary action by performing spatial mapping and object detection in advance before actual manipulation tasks. The system pre-processes the environment data, identifies objects of interest, and prepares manipulation plans, so that when objects need to be manipulated, the processing is already complete or near-complete
Solution Approach 2:
The system implements continuity of useful action by continuously capturing and processing image data as the sensor moves through space. Rather than stopping to process batches of data, the system maintains continuous data flow and processing, efficiently utilizing the movement to gather information that is immediately analyzed
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables accurate and efficient identification and picking of objects across a large spatial area, improving the reliability and precision of object manipulation by considering a greater volume of image data and providing reliable positioning for gripping components.
Implementation Method 1
In one embodiment, the 3D image sensing device is a LIDAR device
Data Source
AI summary
A LIDAR device is positioned on a crane and is moved along a track to collect 3D image data of the area below the crane. The resulting image data is sent to a computing device that applies or more filtering algorithms to the image data and searches for known image shapes therein through a comparison of the image data and one or more 3D object models based on known shapes or geometric primitives. If an object is identified in the image data and is determined to be accessible, position information for that object may be sent to a device configured to control movement of the crane to grab/pick up the object.