Deep Learning LED Recognition for Indoor Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional indoor positioning methods using LED lights are inefficient due to high costs, complex hardware requirements, and low precision, making them unsuitable for large-scale applications.
Innovation Solution
An LED light source recognition method based on deep learning that uses a CMOS camera to detect dark stripes, preprocess the images, and determine light source feature encoding sequences to identify LED lights, allowing for efficient indoor positioning without additional hardware or complex calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional indoor positioning methods (WiFi fingerprint matching, Bluetooth beacons, RFID readers, ultrasonic transceivers) are used, then positioning functionality is achieved, but deployment cost and device complexity increase significantly
Solution Approach 1:
The patent reuses existing LED lighting infrastructure for dual purposes: illumination and positioning beacons. By encoding positioning information in LED light emissions (through color changes, pulsing patterns, or modulation), the system eliminates the need for separate positioning hardware like Bluetooth beacons or RFID readers, thereby reducing deployment complexity while maintaining positioning functionality
Solution Approach 2:
The system uses the already-deployed LED lighting infrastructure to provide positioning services. The existing LED lights serve themselves as positioning beacons by modulating their light output with encoding information, eliminating the need for additional dedicated positioning devices and reducing overall system complexity
2Measurement precision
If multilateral positioning methods (TOA, TDOA, AOA, RSS) are used with LED lights, then positioning precision is improved, but hardware complexity and calibration requirements increase
Solution Approach 1:
The patent replaces complex hardware-based multilateral positioning systems (requiring antenna arrays, precise time synchronization, and specialized receivers) with an image-based optical system. A standard camera or image sensor captures LED light patterns, and computer vision algorithms extract positioning information, thereby reducing receiver hardware complexity while maintaining precision
Solution Approach 2:
Instead of directly measuring physical parameters like time of arrival or signal strength with specialized hardware, the system creates an optical copy of the LED emissions through image capture. The camera records the spatial and temporal patterns of LED lights, which are then decoded to determine position, simplifying the receiver requirements
3Reliability
If image sensor imaging method is used to estimate receiver position, then positioning functionality is achieved, but computational complexity and hardware requirements increase
Solution Approach 1:
The patent segments the visual field into discrete LED light sources, each identifiable by unique characteristics (position, color, pulsing pattern). By detecting and tracking individual LED emitters rather than processing the entire scene continuously, the computational burden is reduced while maintaining positioning functionality
Solution Approach 2:
The system uses simplified image processing that focuses only on detecting LED light patterns rather than performing full scene understanding or complex 3D reconstruction. By applying partial processing (detecting only relevant LED features) rather than exhaustive analysis, computational complexity is reduced while achieving sufficient positioning accuracy
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
Enables precise and cost-effective indoor positioning using LED lights, reducing computational complexity and power consumption, and is suitable for real-time applications in complex indoor environments.
Implementation Method 1
In an LED lighting environment, a spline frame is obtained through a CMOS camera
Data Source
AI summary
The disclosure provides an LED light source recognition method, device, apparatus, and medium based on deep learning, belonging to the field of indoor positioning and navigation technology. The method includes the following. In an LED lighting environment, a spline frame is obtained through a CMOS camera, in which the spline frame is an image having dark stripes. The spline frame is input to a target detection model, a dark stripe detection result output by the target detection model is obtained. The multiple rectangular boxes are preprocessed, and based on the predicted classification corresponding to each preprocessed rectangular box, an image feature encoding sequence of the spline frame is determined. The image feature encoding sequence is compared with a light source feature encoding sequence corresponding to each LED light source to determine the light source feature encoding sequence that best matches the spline frame.