People Counting Depth Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for counting people in a scene, such as using visible spectrum and thermal cameras, face inaccuracies due to shadows and temperature-related issues, which affect detection rates and computational complexity.
Innovation Solution
A depth sensor-based system that obtains depth data to discern foreground objects from background, matches them with a reference model of a human head, and combines this information with visible spectrum camera data to improve detection accuracy and reduce complexity through single-scale detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visible spectrum cameras are used to detect objects in a scene, then the detection system can capture visual information, but detection accuracy deteriorates when shadows present similar shapes and sizes to actual objects
Solution Approach 1:
The patent introduces thermal camera data as an intermediary to transfer or perform the detection action. By fusing visible spectrum camera data with thermal camera data, the system uses thermal information as a mediator to distinguish objects from shadows, since shadows do not have thermal signatures while actual objects do.
Solution Approach 2:
The patent combines data from two different sensing modalities (visible spectrum and thermal) to create a composite detection result. This multi-modal data fusion approach leverages the complementary strengths of both sensor types to overcome the limitations of using either sensor alone.
2Measurement precision
If thermal cameras are used to detect motion in a scene, then the detection system can identify moving objects, but detection accuracy deteriorates when object temperatures are the same as or near to the temperature of the moving objects
Solution Approach 1:
The patent introduces visible spectrum camera data as an intermediary to transfer or perform the detection action. By fusing thermal camera data with visible spectrum camera data, the system uses visual information as a mediator to detect objects that have similar temperatures, since visual features remain distinguishable even when thermal signatures are similar.
Solution Approach 2:
The patent combines data from two different sensing modalities (thermal and visible spectrum) to create a composite detection result. This multi-modal data fusion approach leverages the complementary strengths of both sensor types to overcome the limitations of using either sensor alone.
3Adaptability or versatility
If multiple scales detector is applied for object detection, then the detection system can detect objects of various sizes, but computational complexity increases
Solution Approach 1:
The patent changes the detection parameters by using fixed-size sliding windows instead of multiple scales. The system adjusts the position and size of detection windows based on depth information and object priors, achieving adaptability without requiring multiple scale detectors, thus reducing computational complexity.
Solution Approach 2:
The patent performs preliminary filtering using depth information to identify potential object regions before applying the detector. By pre-processing the data to locate promising detection regions, the system reduces the search space and computational burden of the subsequent detection step.
Data Source
AI summary
A sensor system according to an embodiment of the invention may process depth data and visible light data for a more accurate detection. Depth data assists where visible light images are susceptible to false positives. Visible light images (or video) may similarly enhance conclusions drawn from depth data alone. Detections may be object-based or defined with the context of a target object. Depending on the target object, the types of detections may vary to include motion and behavior. Applications of the described sensor system include motion guided interfaces where users may interact with one or more systems through gestures. The sensor system described may also be applied to counting systems, surveillance systems, polling systems, retail store analytics, or the like.