Computer Vision Safety Monitoring for Machine Proximity Violations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems lack the capability to accurately and timely detect unsafe behaviors around mechanical equipment, such as conveyor belts, leading to potential injuries and increased workplace safety risks.
Innovation Solution
A computer-implemented method using computer vision that accesses video data from image sensors to detect people and motion status of machine components, executes object detection and pose estimation models to identify safety rule violations, and transmits alerts or control signals to mitigate risks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If computer vision systems are deployed to detect unsafe behaviors in real-time, then workplace safety is improved, but system complexity and implementation cost increase
Solution Approach 1:
The system segments the safety monitoring task into distinct functional modules: video data acquisition from multiple sensors, object detection for identifying people and equipment, pose estimation for analyzing body positions, motion status detection for tracking component movement, and safety rule violation detection. Each module processes specific aspects of safety monitoring independently, improving overall system reliability while managing complexity through modular architecture.
Solution Approach 2:
The patent introduces an intermediary processing layer that bridges raw video data and safety decisions. This layer includes frame sampling mechanisms, pixel difference calculation systems, and intermediate detection models that transform raw visual information into structured safety assessments. The intermediary layer buffers and processes data flow, reducing the direct complexity between sensors and control systems.
2Measurement precision
If multiple video frames are sampled and compared at high frame rates to accurately detect motion status, then detection precision is improved, but processing time and computational load increase
Solution Approach 1:
The system implements periodic frame sampling at predetermined frame rates rather than continuous processing. Video frames are sampled at specific intervals, and motion detection is performed periodically by comparing successive sampled frames. This periodic approach maintains adequate motion detection precision while significantly reducing computational load and processing time compared to continuous frame-by-frame analysis.
Solution Approach 2:
The patent applies partial action by comparing only specific regions of video frames using pre-defined polygon regions of interest. Instead of processing entire frames, the system focuses computational resources on relevant areas where motion is most likely to occur, such as regions containing people or critical equipment components. This selective processing maintains detection precision for critical areas while reducing overall processing time.
3Measurement precision
If pixel difference threshold is set low to improve motion detection sensitivity, then false positives increase, but if set high to reduce false positives, then detection sensitivity decreases
Solution Approach 1:
The system employs feedback mechanisms where detection results from multiple frames and multiple detection models are aggregated and evaluated. Motion detection decisions are not based on single frame comparisons but on accumulated evidence from sequential frame analyses. The system adjusts its response based on the consistency and persistence of detected motion patterns, allowing low thresholds for sensitivity while using temporal feedback to filter out transient false positives.
Solution Approach 2:
The patent implements preliminary filtering actions before final motion detection decisions. Pre-defined polygon regions of interest are established in advance to focus detection on critical areas. Motion patterns are pre-analyzed through multiple frames before triggering alerts, and detection models are pre-trained on safety-relevant patterns. This preliminary preparation enables sensitive detection while reducing false positives through pre-established criteria.
Data Source
AI summary
In some examples, a system can access video data collected from one or more image sensors, the video data showing a region of interest proximate to a machine. The system can execute an object detection model to detect that a person is within the region of interest proximate to the machine based on the video data. The system can detect a motion status of a component of the machine. The system can execute a pose estimation model on the video data to estimate a pose of the person with respect to the machine. The system can detect a safety rule violation based on the pose of the person with respect to the machine, and the motion status of the machine. The system can transmit a signal to a controller of the machine in response to detecting the safety rule violation.