Coded Light Localization Using Rolling-Shutter Fourier Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing localization systems using coded light for mobile terminals face complex processing, high energy consumption, and susceptibility to errors due to supply electronics and lack of false recognition exclusion methods, especially when multiple light sources are present.
Innovation Solution
A localization method using analogue periodic signals with Fourier transform processing to identify light sources by analyzing regions of interest, reducing processing load and energy consumption through sparse sampling and compatibility checks, and eliminating false recognitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If complex processing techniques (Gabor filter, M-fold Fourier transform) are used to recognize identification codes, then the system can distinguish multiple light sources, but processing complexity and energy consumption increase significantly
Solution Approach 1:
The patent segments the image into multiple blocks and processes each block independently to identify light sources. This division reduces the computational complexity of analyzing the entire image at once, while still maintaining the ability to distinguish multiple closely-spaced light sources through localized processing.
Solution Approach 2:
The patent extracts only the relevant features from each image block (temporal modulation patterns) rather than processing all pixel data. By focusing extraction on the specific characteristics needed for identification code recognition, the system reduces processing load while maintaining reliability.
2Reliability
If complex processing techniques are applied to recognize identification codes, then source identification capability is enhanced, but energy consumption increases
Solution Approach 1:
The patent applies processing only to selected image blocks that contain light sources rather than processing the entire image. This partial action approach reduces energy consumption by avoiding unnecessary processing of empty or irrelevant regions, while still achieving accurate code recognition where needed.
3Reliability
If image subdivision into predetermined blocks is performed to distinguish multiple light sources, then source separation capability is improved, but errors increase when sources are close together
Solution Approach 1:
The patent dynamically adjusts the block size and processing parameters based on the detected light source characteristics and their spatial relationships. When sources are close together, the system adapts its processing strategy to maintain precision, rather than using fixed predetermined blocks that may cause errors.
4Productivity
If standard Fourier transform processing is used, then identification code recognition is achieved, but susceptibility to errors from supply electronics increases
Solution Approach 1:
The patent performs preliminary filtering and preprocessing of the temporal modulation signals before applying Fourier transform. This preliminary action removes noise and artifacts introduced by supply electronics, ensuring that the subsequent processing operates on cleaner data and reduces susceptibility to errors.
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
Simplifies processing and reduces energy consumption by analyzing a single image, minimizing errors through fixed length profiles and compatibility checks, while maintaining accurate localization.
Implementation Method 1
a common technique for the acquisition of images of coded light sources by the cameras of mobile terminals is the so called 'rolling shutter' technique, according to which the different lines of the matrix of pixels of the camera are read and converted into images at different times, so that the temporal modulation of the source is converted into a spatial modulation in the image acquired
Implementation Method 2
The system detects the codes form the spatial pattern of modulations (i.e. from the light distribution in the different lines) by using a filter, more particularly a Gabor filter
Implementation Method 3
The processing includes an M-fold application (M = number of columns) of a one-dimensional discrete Fourier transform and of a two-dimensional Fourier transform to each of the regions into which each image is divided
Data Source
Figure 1~3(b)
Figure 4
Figure 5~6
AI summary
A method of localising a user inside a space (1) in which a plurality of sources (2a...2n) of coded visible light are located includes the steps of: modulating each source (2a...2n) with at least one periodic signal (5a...5n) and assigning an identification code related to the frequency or the frequencies of the at least one periodic signal to the source (2a...2n); acquiring (103) an image of a scene in which one or more sources (2a...2n) are present by means of a camera (30) including a matrix of pixels sequentially read according to a rolling shutter technique; processing (104, 106) the image in order to recognise the identification code of the or each source (2a...2n); and accessing (107), by using the identification code recognised, information associating the codes with the sources (2a...2n) and their positions in the space (1). The processing step (104, 106) includes the steps of: identifying (104, 202 - 204), in the image of the scene, one or more regions of interest (40), each substantially comprising the only block of pixels containing a source (2a...2n); and for a region of interest (40) compatible with the system (10): extracting (302) a brightness profile of the region of interest (40) in the reading direction of the matrix of pixels; bringing (303, 304) said brightness profile to a pre-set and constant length; and recognising (305, 306) the identification code by submitting the brightness profile to a transformation from the space domain to the wave number domain.