Adaptive Infrared Projection Control for Depth Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current depth estimation methods in computer vision, such as passive and active stereo vision, face challenges in reducing power consumption and ensuring eye safety during infrared (IR) projection, as they typically require continuous IR light emission regardless of the scene or user conditions.
Innovation Solution
Implementing adaptive IR projection control by using a processor to detect regions of interest (ROI) in images and dynamically adjust the IR projector's power and timing based on conditions such as facial recognition, eye state, distance, environmental brightness, and battery power, allowing for reduced IR projection time and power usage while enhancing eye safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If continuous IR light emission is used for depth estimation, then measurement precision is improved, but use of energy increases and eye safety deteriorates
Solution Approach 1:
The system implements periodic IR projection controlled by a state machine that cycles through different projection modes (full projection, reduced projection, or no projection) based on detected conditions. The IR projector is activated only during specific time windows when depth data is needed and conditions permit, rather than operating continuously, thereby reducing energy consumption while maintaining measurement capability.
Solution Approach 2:
The system dynamically adjusts IR projection parameters including power level and duration based on real-time conditions such as detected facial features, eye state, distance to subject, and ambient light levels. The state machine transitions between operational states to optimize projection settings, enabling the system to adapt power consumption to actual measurement needs rather than using fixed continuous projection.
2Measurement precision
If continuous IR light emission is used for depth estimation, then measurement precision is improved, but object-affected harmful factors increase
Solution Approach 1:
The state machine controls IR projection to occur only during specific periodic time windows when depth data acquisition is required, with intervals where projection is reduced or stopped. This periodic operation reduces cumulative eye exposure to IR light while maintaining sufficient data collection for depth estimation, thereby improving eye safety without completely sacrificing measurement precision.
Solution Approach 2:
The system dynamically adjusts projection power and duration based on real-time detection of eye state, facial distance, and other safety-relevant conditions. When eyes are detected or when subjects are close, the system reduces or stops projection to minimize harmful exposure. This dynamic adaptation allows the system to maintain measurement precision when safe while protecting eyes when risks are present.
3Use of energy by moving object
If adaptive IR projection control is implemented, then use of energy is reduced, but device complexity increases
Solution Approach 1:
The control system is segmented into discrete functional modules: image sensor for condition detection, processor for analyzing detected features, state machine for decision logic, and IR projector for light emission. Each module performs a specific function, making the overall complex system manageable through modular design. The state machine uses predefined states and transitions based on simple condition checks, reducing the complexity burden of adaptive control.
Solution Approach 2:
The system uses its own image sensor and processor to detect conditions and automatically adjust IR projection parameters without external intervention. The state machine autonomously transitions between operational states based on detected conditions, enabling the system to self-regulate power consumption based on actual needs rather than requiring manual control or complex external management systems.
Data Source
AI summary
A processor or control circuit of an apparatus receives data of an image based on sensing by one or more image sensors. The processor or control circuit also detects a region of interest (ROI) in the image. The processor or control circuit then adaptively controls a light projector with respect to projecting light toward the ROI.