Smoke Detection Imaging Control for Repeated Cloud Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing smoke detection systems face inefficiencies and inaccuracies due to repeated detection of the same smoke and failure to detect multiple smoke clouds, leading to decreased efficiency and accuracy.
Innovation Solution
A system utilizing an object detection device with a moveable component and imaging components to determine if subjects in current and previous images belong to the same object, allowing the system to move the imaging component to avoid repeated detection and ensure comprehensive smoke detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the smoke detection device stops moving to perform fire detection when smoke is detected, then the detection accuracy is improved, but the device cannot move and repeatedly detects the same smoke, decreasing detection efficiency
Solution Approach 1:
The system uses feedback from image recognition results and position information to control movement. When smoke is detected, the device continues moving and uses image recognition to determine if the detected smoke is new or previously detected. Based on this feedback, the device intelligently decides whether to continue moving or stop for detailed detection, resolving the contradiction between maintaining movement for efficiency and stopping for accuracy.
Solution Approach 2:
The device dynamically adjusts its movement state based on real-time detection results. Instead of a fixed stop-when-detected rule, the system dynamically determines whether to continue moving or stop by comparing current detection results with historical position information and image recognition data, optimizing both efficiency and accuracy adaptively.
2Measurement precision
If the smoke detection device remains stationary to detect smoke accurately, then detection precision is improved, but multiple smoke clouds in close regions cannot be fully detected
Solution Approach 1:
The device dynamically adjusts its movement state based on real-time detection results. When multiple smoke clouds are present, the system continues moving to capture different viewing angles and detect additional smoke clouds, rather than remaining stationary. This dynamic approach ensures both accurate detection of individual clouds and comprehensive detection of multiple clouds in the region.
Solution Approach 2:
The system adds the spatial dimension of movement to the detection process. By continuing to move and capture images from different positions and angles, the device can detect multiple smoke clouds that exist in close regions, transforming a single-point detection limitation into a multi-position detection capability.
3Area of stationary object
If the device continuously moves to detect multiple smoke clouds, then detection coverage is improved, but repeated detection of the same smoke increases, decreasing overall efficiency
Solution Approach 1:
The system uses feedback from image recognition and position comparison to identify repeated detections. When the device detects smoke at a new position, it uses image recognition to determine if this is a new smoke cloud or a previously detected one. This feedback mechanism allows the device to continue moving for broad coverage while filtering out repeated detections of the same smoke, maintaining efficiency.
Solution Approach 2:
Image recognition acts as an intermediary between movement and detection results. It processes the relationship between current and historical detection data, determining whether detected smoke represents a new target or a repeat detection. This intermediary enables the system to maintain continuous movement while accurately distinguishing between new and repeated smoke detections.
Data Source
AI summary
Systems and methods for smoke detection using object detection devices are provided. The object detection device may include a first imaging component and a moveable component configured to move the first imaging component. The system may obtain a current image of a first subject acquired by the first imaging component at a current time. The system may determine whether the first subject and a second subject among one or more second subjects belong to a same object. In response to determining that the first subject and the one of the one or more second subjects belong to the same object, the system may cause the moveable component to move the first imaging component.