Adaptive Drone Perception Using Dynamic Camera Parameter Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Movable apparatuses such as drones require artificial intelligence systems that can process information quickly, recognize distant objects, and operate with low power consumption while maintaining accurate object detection and tracking.
Innovation Solution
An electronic device equipped with a neural processing unit (NPU) and signal generator to control cameras, adjusting parameters like viewing angle, focal distance, and image resolution based on flight conditions to enhance object detection and tracking accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the drone uses high processing speed AI system to detect objects quickly, then the object detection speed improves, but the power consumption increases
Solution Approach 1:
The patent implements dynamic adjustment of camera parameters (focal length, aperture, exposure time) based on real-time flight conditions and object detection requirements. This allows the system to optimize processing speed and power consumption adaptively - using higher processing speeds only when necessary for quick decision-making during critical flight phases, while reducing power consumption during steady-state operation.
Solution Approach 2:
The system changes operational parameters dynamically - adjusting camera focal length, aperture size, and exposure time based on distance to target and lighting conditions. This enables the AI system to maintain high detection accuracy while consuming less power by using optimal parameter combinations rather than always operating at maximum performance levels.
2Measurement precision
If the drone uses high-resolution cameras to recognize distant objects, then the object recognition accuracy improves, but the power consumption increases
Solution Approach 1:
The patent dynamically adjusts camera resolution and focal length based on the distance to the detected object and the required level of detail. When objects are far away, the system increases focal length and uses higher resolution modes. When objects are closer or during routine monitoring, the system reduces resolution settings, thereby maintaining recognition accuracy when needed while minimizing power consumption during normal operation.
3Duration of action of moving object
If the drone extends flight duration with low power mode, then the flight time increases, but the object detection accuracy decreases
Solution Approach 1:
The system implements periodic high-accuracy detection cycles interspersed with lower-power monitoring phases. During extended flight, the drone performs comprehensive object detection and parameter optimization at regular intervals, then transitions to energy-efficient monitoring mode between these cycles. This periodic approach ensures detection accuracy is maintained when needed while allowing extended flight time through reduced power consumption during intermediate periods.
4Measurement precision
If the drone adjusts camera parameters dynamically to improve detection accuracy, then the object detection accuracy improves, but the system complexity increases
Solution Approach 1:
The patent implements feedback loops where the AI system continuously monitors detection results, flight conditions, and environmental factors, then automatically adjusts camera parameters in response. This feedback mechanism enables the system to achieve high detection accuracy through adaptive parameter tuning without requiring complex manual control systems, as the automated feedback-driven adjustments simplify the overall system architecture while maintaining precision.
Data Source
AI summary
An electronic device mounted on a fixed or a movable apparatusapparatus is provided. The electronic device may comprise an image signal processor (ISP) for at least one camera; a neural processing unit (NPU), including a plurality of processing elements (PEs), configured to: process an operation of an artificial neural network model trained to detect or track at least one object, based on an input feature map generated from at least one image, which is acquired via the ISP from the at least one camera, and output an inference result; and a signal generator generating a signal applicable to the at least one camera or the ISP.