People Counting Near Windowed Doors Using Shape Approximation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for counting people using video cameras near external windowed doors struggle with accurately distinguishing between real objects and spurious light, which can cause counting errors due to dynamic light reflections from door surfaces and other sources.
Innovation Solution
A system that captures video images using a camera with two-dimensional shapes on the image plane, where a computer system detects objects by approximating shapes such as ellipses to differentiate between real objects and spurious light, considering shape properties like size and compactness, and determines the direction of movement to accurately count objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video cameras are used to count people near external windowed doors, then people counting capability is provided, but counting accuracy deteriorates due to spurious light from door reflections and ambient light deflection
Solution Approach 1:
The patent segments the detection task into multiple stages: first detecting motion in the video feed, then analyzing the motion characteristics to distinguish between people and spurious light. The system divides the detection process into initial motion detection, shape analysis, and final classification, allowing accurate counting despite light interference
Solution Approach 2:
The patent changes detection parameters dynamically by adjusting sensitivity thresholds based on environmental conditions. The system monitors light conditions and adjusts detection parameters accordingly, switching between different detection modes to maintain accuracy in varying lighting conditions near external doors
2Productivity
If motion detection is used to identify people, then people detection capability is provided, but false positives increase due to dynamic light blobs from swinging doors
Solution Approach 1:
The patent applies dynamics by making the detection system adaptive and responsive to changing conditions. The system dynamically adjusts detection parameters based on real-time environmental conditions, and uses temporal analysis of motion patterns to distinguish between reliable detections and false positives from dynamic light sources
Solution Approach 2:
The system incorporates feedback mechanisms where detection results are continuously refined based on accumulated data. The patent uses feedback from multiple detection cycles to improve accuracy, adjusting detection thresholds and parameters based on learned patterns of genuine motion versus spurious light effects
3Measurement precision
If shape approximation is used to differentiate objects, then object discrimination capability is provided, but computational complexity increases
Solution Approach 1:
The patent applies partial action by implementing a multi-stage filtering process where only candidates meeting initial simple criteria undergo more complex shape analysis. The system performs basic motion detection first, then applies sophisticated shape approximation only to promising candidates, reducing overall computational complexity while maintaining high discrimination accuracy
Data Source
AI summary
A system for counting objects, such as people, crossing an area includes a camera configured for capturing video images along a surface in the area. The surface includes a plurality of detectable features on the surface. A user interface allows an individual to define a region where object detection is desired. A processor processes the images and counts a detected object in the region when criteria for counting are satisfied.