Auto-Calibrating 3D Sensor for Elevator Installation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The calibration of 3D sensors in passenger conveyance systems, such as elevators, is labor-intensive and inconvenient, requiring significant effort and specialized targets, especially when determining extrinsic parameters like pitch, yaw, and roll, which are often ambiguous due to non-uniform object movement and lack of absolute size references.
Innovation Solution
An auto-calibration system that automatically determines the pitch of a 3D sensor by using a fiducial mark, inclinometer, or projector to emit light patterns, allowing for minimal installer effort and eliminating the need for special calibration targets by calculating the pitch using measured distances and angles, thereby aligning the sensor coordinate system with the world coordinate system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional manual calibration methods are used with portable 3D calibration jigs and multiple measurement points, then calibration accuracy is improved, but installation labor and time requirements increase significantly
Solution Approach 1:
The system performs self-calibration by automatically detecting the floor plane and calculating extrinsic parameters using processor-based computation. The depth sensor and image sensor work together to automatically determine calibration parameters without requiring manual placement of calibration objects or complex measurement procedures, enabling the system to calibrate itself rapidly during installation.
Solution Approach 2:
The patent replaces manual mechanical calibration procedures with automated optical and computational methods. Instead of physically positioning calibration jigs and manually recording coordinates, the system uses image sensors and depth sensors to capture spatial information and computationally derives calibration parameters, substituting mechanical operations with optical-field and information-processing operations.
2Reliability
If manual calibration procedures are used requiring placement of detectable objects at multiple locations, then complete calibration data is obtained, but installer effort and complexity increase
Solution Approach 1:
The system automatically performs the complete calibration sequence without installer intervention. The processor executes calibration algorithms that utilize data from the depth sensor and image sensor to compute all necessary extrinsic parameters, eliminating the need for installers to manually position objects, take measurements, and enter data into calibration software.
Solution Approach 2:
The calibration system uses a multi-functional approach where the same sensor suite (depth sensor and image sensor) performs both operational functions and calibration functions. This universal system can simultaneously capture spatial data for both normal operation and self-calibration, eliminating the need for separate calibration equipment and procedures.
3Ease of manufacture
If 3D depth sensing systems use sensor-centric coordinate systems, then sensor data acquisition is simplified, but alignment with world coordinate systems becomes complex
Solution Approach 1:
The patent replaces complex coordinate transformation mechanics with computational algorithms. The processor automatically computes the transformation between sensor-centric and world coordinate systems using spatial relationships derived from depth and image sensor data, substituting manual coordinate alignment procedures with automated information processing.
Solution Approach 2:
The system introduces an intermediary computational layer that mediates between the sensor-centric coordinate system and the world coordinate system. The processor calculates transformation parameters and applies coordinate transformations as an intermediary step, enabling seamless integration of sensor data into the world coordinate framework without direct manual alignment.
Data Source
Figure 1
Figure 2~3
Figure 4~6
AI summary
An installation process for a sensor associated with a passenger conveyance system, the process including at least partially automatically calibrating a sensor coordinate system to a world coordinate system via a calibration matrix, wherein the sensor coordinate system is at least partially obtained using a depth map.