TOF Camera Threshold Adaptation for Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing person position detection apparatuses using the TOF method face challenges in setting optimal threshold values, leading to either excessive shot noise or partial object disappearance, especially in moving object scenarios, as users must manually adjust thresholds, increasing their burden.
Innovation Solution
The apparatus automatically and adaptively sets the threshold value based on the positional relation between the TOF camera and the object, using an installation angle detection unit and a threshold value storage unit to optimize background removal in distance images, ensuring clear images with minimal noise even in dynamic environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the threshold value for determining the distance displacement region is set to be small, then the shot noise is reduced, but a lot of shot noise is output and erroneously recognized as the distance displacement region
Solution Approach 1:
The threshold value is made dynamic rather than fixed. The apparatus automatically adjusts the threshold value based on the installation angle of the TOF camera and the distance to the object, allowing the threshold to adapt to different measurement conditions and optimize the balance between noise reduction and object detection accuracy
Solution Approach 2:
The threshold value is changed as a variable parameter that depends on installation angle and object distance. By storing threshold values in a lookup table indexed by these parameters, the system optimizes noise filtering performance for each specific measurement scenario without requiring manual user adjustment
2Object-generated harmful factors
If the threshold value is set to be large, then the shot noise is reduced, but a part or the entirety of the object shape disappears
Solution Approach 1:
The threshold value dynamically adapts to the specific measurement conditions (installation angle and object distance), ensuring that it is high enough to filter noise but low enough to preserve object shapes in each scenario, eliminating the need for a single fixed threshold that works for all conditions
Solution Approach 2:
The threshold parameter is changed based on the relationship between camera installation angle and object distance. The lookup table stores optimized threshold values for different parameter combinations, allowing the system to maintain optimal noise filtering while preserving object geometry across diverse measurement scenarios
3Measurement precision
If the threshold value is manually adjusted by the user, then the optimal threshold can be set, but the burden of the user increases in a moving object
Solution Approach 1:
The apparatus performs self-optimization of the threshold value without requiring user intervention. By automatically detecting the installation angle and object distance, and retrieving the appropriate threshold from the lookup table, the system serves itself in optimizing noise filtering parameters, making the system easy to operate even in moving object scenarios
Solution Approach 2:
The system uses feedback from the installation angle detection and object distance measurement to automatically adjust the threshold value. This closed-loop approach ensures the optimal threshold is selected based on actual measurement conditions, eliminating the need for manual user adjustment while maintaining high measurement precision
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables the removal of shot noise and accurate detection of person positions and postures by dynamically adjusting the threshold value according to the camera's installation angle, object distance, and posture, resulting in a clear and noise-reduced image output.
Implementation Method 1
A technique of measuring a distance to an object on the basis of a light transmission time (hereinafter, referred to as a TOF method: Time Of Flight)
Data Source
Figure 1
Figure 2A~2B
Figure 3
AI summary
An object position detection apparatus 1 includes a distance image generation unit 10 which generates a two-dimensional distance image by measuring a distance to a subject including an object on the basis of a light transmission time, an object image generation unit 16 which generates an object image in which the object is extracted from the distance image of the subject, and an installation angle detection unit 20 which detects an installation angle of the distance image generation unit 10. The object image generation unit 16 includes a differentiator 17 which performs a process of removing a background other than the object from the distance image of the subject and a threshold value d for removing the background in the differentiator 17 is set in response to the installation angle of the distance image generation unit 10 detected by the installation angle detection unit 20.