Maritime Obstacle Detection with Adaptive Image Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face challenges in accurately acquiring object information from images captured in environments with high levels of environmental noise, such as fog or rain, which can hinder the performance of artificial neural networks used in maritime settings.
Innovation Solution
A method utilizing two artificial neural networks: one for processing images with low noise levels to extract object information related to distance and type, and another that reduces noise levels in images using a noise-reducing technique, allowing for accurate object information acquisition even in noisy conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single artificial neural network is used to process images in maritime environments, then the system complexity is low, but the measurement precision of object information degrades when environmental noise is high
Solution Approach 1:
The system segments the image processing task into two distinct neural networks: a first neural network that processes images with low noise levels directly, and a second neural network that reduces noise in images with high environmental noise before processing. This segmentation allows each network to be optimized for specific noise conditions, improving overall measurement precision without requiring a single overly complex system to handle all scenarios.
Solution Approach 2:
The system dynamically selects which neural network to use based on the detected noise level in the input image. When environmental noise is detected to be below a threshold, the first neural network is used; when noise exceeds the threshold, the second neural network is invoked for noise reduction first. This dynamic adaptation optimizes processing accuracy for varying conditions while maintaining reasonable system complexity.
2Measurement precision
If noise reduction processing is applied to all images, then object information accuracy is maintained, but processing time increases
Solution Approach 1:
Instead of applying noise reduction processing to all images, the system applies it only partially - specifically to images where environmental noise exceeds a predetermined threshold. For images with acceptable noise levels, the system skips the noise reduction step and proceeds directly to object detection. This partial application of processing maintains accuracy where needed while minimizing unnecessary processing time for already-clear images.
Data Source
AI summary
The present invention relates to a method for acquiring an object information, the method comprising: obtaining an input image acquired by capturing a sea; obtaining a noise level of the input image; when the noise level indicates a noise lower than a predetermined level, acquiring an object information related to an obstacle included in the input image from the input image by using a first artificial neural network, and when the noise level indicates a noise higher than the predetermined level, obtaining a noise-reduced image of which the environmental noise is reduced from the input image by using a second artificial neural network, and acquiring an object information related to an obstacle included in the sea from the noise-reduced image by using the first artificial neural network.