AI Adaptive Optics Without Deformable Mirror
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive optical systems require expensive deformable mirrors to correct wave front distortions, limiting their accessibility and increasing costs.
Innovation Solution
An adaptive optical system using artificial intelligence that eliminates the need for a deformable mirror by employing a reflector, beam splitter, camera, wave front sensor, and correction processor to generate and correct images based on wave front information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a deformable mirror is used to correct wave front distortions in adaptive optical systems, then image quality and resolution are improved, but system cost and complexity increase significantly
Solution Approach 1:
The patent replaces the mechanical deformable mirror with a computational approach using neural networks. Instead of physically deforming a mirror to correct wave front distortions, the system uses a neural network to learn and correct distortion patterns from captured images, substituting mechanical adjustment with intelligent computation.
Solution Approach 2:
The patent creates a virtual copy of the distortion correction process through neural network training. The neural network learns distortion characteristics from training data and generates corrected images by applying learned transformations, effectively copying the correction function without requiring physical deformable components.
2Reliability
If a deformable mirror is used to actively change the reflector surface, then wave front distortion is corrected, but the system becomes more expensive and complex
Solution Approach 1:
The patent replaces the mechanical deformable mirror with a computational neural network that processes images to correct distortions. The neural network receives distorted images as input and outputs corrected images, eliminating the need for mechanical surface adjustment while maintaining distortion correction reliability.
Solution Approach 2:
The neural network acts as an intermediary between the captured image and the final corrected image. Instead of directly modifying the optical path with a deformable mirror, the neural network processes the image data to remove distortions, serving as a computational mediator that achieves the same correction effect.
3Manufacturing precision
If exposure time is increased to collect weak incident light for sharp images, then image sharpness is improved, but the system becomes more susceptible to atmospheric errors
Solution Approach 1:
The patent employs feedback mechanisms where the neural network continuously learns from and adapts to atmospheric distortion patterns. The system captures images with longer exposure times to maintain sharpness, then uses the neural network to feedback-correct the atmospheric errors introduced during capture, effectively compensating for the increased vulnerability to atmospheric conditions.
Solution Approach 2:
The neural network performs preliminary learning of distortion patterns during training, preparing correction models before actual image capture. This preliminary action enables the system to accurately compensate for atmospheric errors even when using longer exposure times, as the correction algorithms are pre-trained on various atmospheric conditions.
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
This system effectively corrects image distortions without a deformable mirror, reducing costs and enabling the generation of high-quality images with minimal distortion across various applications.
Implementation Method 1
a reflector reflecting incident light to one side
Implementation Method 2
a beam splitter splitting light incident from the reflector into a first direction and a second direction
Implementation Method 3
at least one wave front sensor positioned in the second direction with respect to the beam splitter, detecting a wave front of incident light
Data Source
AI summary
The present invention relates to an adaptive optical system using artificial intelligence and not having a deformable mirror, the adaptive optical system being capable of generating an image with no distortion or minimized distortions, without using a surface-deformable reflector which is relatively expensive equipment, and also being capable of generating a high-quality image.