Optimizing Embedded Industrial Safety Systems Based on Time-of-flight Depth Imaging

Thoman, Peter; Hirsch, A.A.; Wippler, Markus; Hranitzky, Robert

doi:10.1109/escience51609.2021.00048

Cited by 2 publications

(1 citation statement)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Their performance can significantly decrease due to environmental factors such as the internal temperature of the camera, ambient temperature, and humidity, which can affect the accuracy of distance measurements. Additionally, the accuracy of ToF cameras varies significantly with the distance to the target, requiring correction technologies for practical applications [1,2].…”

Section: Introductionmentioning

confidence: 99%

Enhancing ToF Sensor Precision Using 3D Models and Simulation for Vision Inspection in Industrial Mobile Robots

Yang,

Kang,

Eom

2024

Applied Sciences

View full text Add to dashboard Cite

In recent industrial settings, time-of-flight (ToF) cameras have become essential tools in various applications. These cameras provide high-performance 3D measurements without relying on ambient lighting; however, their performance can degrade due to environmental factors such as temperature, humidity, and distance to the target. This study proposes a novel method to enhance the pixel-level sensing accuracy of ToF cameras by obtaining precise depth data labels in real-world environments. By synchronizing 3D simulations with the actual ToF sensor viewpoints, accurate depth values were acquired and utilized to train AI algorithms, thereby improving ToF depth accuracy. This method was validated in industrial environments such as automobile manufacturing, where the introduction of 3D vision systems improved inspection accuracy compared to traditional 2D systems. Additionally, it was confirmed that ToF depth data can be used to correct positional errors in mobile robot manipulators. Experimental results showed that AI-based preprocessing effectively reduced noise and increased the precision of depth data compared to conventional methods. Consequently, ToF camera performance was enhanced, expanding their potential applications in industrial robotics and automated quality inspection. Future research will focus on developing real-time synchronization technology between ToF sensor data and simulation environments, as well as expanding the AI training dataset to achieve even higher accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%