Stripe Image Processing for Visible Light Communication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In visible light communication scenarios, there is an urgent need to balance image reconstruction accuracy and stripe extraction accuracy, as too many black stripes hinder image reconstruction while too few hinder stripe extraction.
Innovation Solution
A stripe image processing method using a generative adversarial network (GAN) is employed, where a dataset of stripe images, stripe sequences, and stripe-free images is used to train both an image reconstruction model and a stripe extraction model, effectively balancing the adversarial tasks of image reconstruction and stripe extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If stripe extraction is optimized to improve communication decoding, then stripe extraction accuracy is improved, but image reconstruction accuracy deteriorates
Solution Approach 1:
The patent divides the processing of stripe images into two separate specialized models: one dedicated to stripe extraction for communication decoding, and another dedicated to image reconstruction for background recovery. This segmentation allows each model to optimize for its specific task without compromising the other, resolving the contradiction between stripe extraction accuracy and image reconstruction accuracy
Solution Approach 2:
The patent introduces a dual-model system as an intermediary between the conflicting requirements of stripe extraction and image reconstruction. The first model extracts stripe sequences for communication purposes, while the second model reconstructs the background image by removing stripes. This intermediary architecture enables both functions to coexist and operate at high accuracy levels simultaneously
2Ease of operation
If too many black stripes are present in the image, then stripe extraction is facilitated, but image reconstruction quality deteriorates
Solution Approach 1:
The patent segments the processing workflow into two distinct pathways: one that leverages abundant stripes for easy stripe extraction and communication decoding, and another that uses a specialized reconstruction model to recover the background image by identifying and removing stripe patterns. This segmentation allows the system to handle images with many stripes effectively for both purposes
3Manufacturing precision
If too few black stripes are present in the image, then image reconstruction quality is improved, but stripe extraction accuracy deteriorates
Solution Approach 1:
The patent creates two specialized processing models that work in parallel: one optimized for detecting even faint stripe patterns in images with few stripes for communication decoding, and another optimized for reconstructing the background image. This segmentation enables the system to maintain high stripe extraction accuracy even when stripes are sparse, while simultaneously achieving high image reconstruction quality
Data Source
AI summary
The disclosure belongs to the field of visible light communication, and specifically discloses a stripe image processing method and apparatus for camera optical communication. The method includes: obtaining a stripe image data set, the stripe image data set including a stripe image sample and a stripe sequence label corresponding to the stripe image sample and a stripe-free image label; based on the stripe image sample and the stripe sequence label corresponding to the stripe image sample and the stripe-free image label, training a generative adversarial network to obtain an image reconstruction model and a stripe extraction model; among them, the image reconstruction model serves as a generator of the generative adversarial network, and the stripe extraction model serves as a discriminator of the generative adversarial network.