Surveillance Software Shadow Elimination via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video-based intelligence surveillance systems face challenges in accurately distinguishing between objects and shadows, particularly in varying lighting conditions, which can lead to false images and reduced detection effectiveness.
Innovation Solution
The system employs software that captures images within a boundary box, adjusts the box's dimensions using vectors and shadow angle calculations to eliminate soft shadows, allowing for accurate detection of human intruders by recalculating the width and height based on shadow presence, thereby improving detection rates and reducing false alarms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If the surveillance system operates in natural light conditions to cover large areas, then the coverage area is improved, but shadow distortion increases making detection unreliable
Solution Approach 1:
The system segments the detected object into two distinct parts: the actual object and its shadow. By analyzing the spatial relationship and geometric characteristics between these segments, the system can identify and exclude shadow regions from detection, thereby maintaining reliable detection across large surveillance areas with natural lighting.
Solution Approach 2:
Instead of trying to detect objects while ignoring shadows, the system inverts the approach by first detecting shadows and then using that information to refine object detection. The shadow detection becomes the primary function that enables subsequent accurate object identification, turning the shadow from a干扰 factor into a useful detection aid.
2Use of energy by moving object
If the system uses ambient light conditions for surveillance, then energy consumption is reduced, but shadow-induced false images increase
Solution Approach 1:
The system converts the harmful effect of shadows into a beneficial detection mechanism. By analyzing shadow characteristics such as geometry, position, and relationship to the light source, the system uses shadows as additional information to improve object detection accuracy rather than treating them merely as noise to be eliminated.
3Measurement precision
If the boundary box includes the entire detected shape, then detection sensitivity is improved, but shadow inclusion increases false alarms
Solution Approach 1:
The boundary box detection is segmented into object-specific boundaries and shadow boundaries. The system calculates separate bounding boxes for each segment, ensuring that the final detection boundary encompasses only the actual object without including shadow regions, thus maintaining high detection sensitivity while reducing false alarms.
Data Source
AI summary
One of the challenges in video surveillance system is the detection of unwanted intruders in secure zones. These zones may be in large facilities such as airports, train stations, port facilities or warehouses. Because many of these locations are outdoors, the presence of shadows is detected on images. In order to effectively eliminate these shadows which will therefore result in a usable figure, these shadows must be eliminated. This is accomplished by forming a contour box around the image and then reducing the shadows based on preset parameters in the software.