Scene-Aware Device Control Using Object Recognition and Context
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current device control methods require user participation or simple automatic triggering, limiting their ability to effectively control devices based on complex scenarios or dynamic environments.
Innovation Solution
A device control method that acquires a scene image, identifies a target object and scene type, and controls associated devices using pre-set strategies by establishing relationships between device images, features, and identifications, enabling automatic control without user intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If automatic triggering is used to control devices, then device control can be achieved without user participation, but the control capability is limited to simple scenarios only
Solution Approach 1:
The system dynamically adapts its control behavior based on the complexity of the scenario. For simple scenarios, it uses automatic triggering with sensor state changes. For complex scenarios, it transitions to scene graph-based reasoning that can handle multiple objects, relationships, and conditional logic, allowing the automation extent to adjust according to scenario requirements
Solution Approach 2:
The control system is segmented into multiple processing levels: basic sensor triggering for simple tasks, scene object identification for medium complexity, and scene graph reasoning for high complexity scenarios. This segmentation allows the system to apply the appropriate level of automation based on the specific scenario requirements
2Measurement precision
If user participation is required to indicate control objects, then device control can be precise, but the operation convenience is reduced
Solution Approach 1:
The system performs self-service by automatically identifying control objects through scene graph construction and reasoning. The scene understanding module autonomously detects objects, relationships, and contextual information, then determines the appropriate control target without requiring user indication, thereby maintaining precision while eliminating user intervention
Solution Approach 2:
The manual user indication process is replaced with an automated scene understanding system that uses image recognition, object detection, and logical reasoning to identify control objects. This substitution maintains or improves identification accuracy while completely eliminating the need for user participation
3Ease of manufacture
If simple sensor state change detection is used, then the control system is easy to implement, but the control capability is limited to fixed devices only
Solution Approach 1:
The scene graph-based control system serves multiple functions: it can handle simple sensor triggering, identify various types of devices through image recognition, understand spatial relationships, and apply different control strategies based on scene context. This universal framework replaces multiple specialized systems with a single adaptable platform
Data Source
AI summary
A device control method, a device control apparatus, an electronic device, a program, and a medium provided by the present disclosure belong to the technical field of computers. The method includes: acquiring a scene image of a target space; acquiring a target scene type matching a target object in the scene image; acquiring a target device associated with the target object shown in the scene image; and controlling the target device according to a device control strategy.