Multiview Depth Calculation

K|Lens Multiview Depth Calculation uses the principles of parallax to extract highly accurate depth information from image data.

K|Lens Multiview Depth Calculation is a state-of-the-art technology that delivers exceptionally accurate 3D information through the clever use of parallax, redundancy and precise calibration.

From parallax to disparity map

The basic principle is based on parallax, an effect that occurs when a scene is viewed from different angles. When recording a scene with several cameras, disparities arise between the images due to parallax. These disparities, i.e. the shifts of pixels between the views, are a direct measure of the distance of objects in the scene. Our Multiview Depth Calculator analyzes these disparities with subpixel accuracy to create a disparity map in which the brightness of each pixel reflects the size of the disparity.

From the disparity map to the depth map

The disparity map serves as the basis for generating the depth map. In combination with the calibration data of the camera system, a depth map is created in which the brightness of each pixel indicates the distance to the camera. Thanks to the individual and sophisticated calibration of our systems, a very high accuracy of the depth map is achieved. Calibration plays a crucial role here, as it forms the basis for the precise conversion of disparities into distances.

Integration of redundancy

A key advantage of our technology lies in its ability to process information from many different perspectives. This redundancy makes the depth calculation extremely robust against disturbances such as reflections and occlusions. Even under difficult lighting conditions or with complex surface structures, reliable disparities and thus depth values can be obtained.

Important technical features

Adaptation to specific applications

Calculating disparity per pixel is a complex process that has different requirements depending on the use case. Therefore, we offer a wide range of algorithms, including traditional, AI-based and hybrid approaches. This diversity makes it possible to select the optimal algorithm for each specific task.

Subpixel precision

The K|Lens Multiview depth calculation supports subpixel precision, which means that disparities are captured with an accuracy that is far below the size of a pixel. This enables extremely fine and precise depth determination.

Highly optimized performance

All algorithms are highly optimized to ensure maximum performance and efficiency. This enables fast and precise depth calculation, even with large amounts of data.

Accelerated processing

Depth calculation benefits from the use of acceleration cards such as graphics cards (GPUs) or HAILO processors. This enables a significant acceleration of calculations and real-time processing.

Wide range of parameters

A wide range of parameters enables fine tuning of the depth calculation to the specific requirements of the application. This makes it possible to optimize the accuracy, speed and robustness of the calculation.

Real-time capabilities

Through optimal parameterization, clock rates of less than one second can be achieved, which predestines the technology for real-time applications such as robotics or inline quality control.

High-resolution depth maps

The technology enables the creation of high-resolution depth maps that allow detailed 3D reconstruction of sophisticated components.

Optimization for specific object geometries

Customer-specific adjustments allow the depth calculation to be optimized for special object geometries. This leads to a further increase in the accuracy and speed of the calculations.

Adaptation to environmental influences

The technology can be adapted to specific environmental influences such as lighting conditions, temperature or vibrations. This improves the robustness of the depth calculation and ensures reliable results even under difficult conditions.

Integration options

Seamless integration with K|Lens Multiview calibration

The software works seamlessly with K|Lens Multiview Calibration, which provides a precise basis for depth calculation from disparities.

Highly specialized algorithms for K|Lens light field sensor data

The software has highly specialized algorithms specifically designed to process data from the K|Lens light field sensor. These algorithms enable the precise calculation of 3D models from the captured light field data.

Adaptability to existing camera arrays

The software is adaptable to existing camera arrays. This allows users to use the software with their existing hardware and optimize their existing workflows.

Data processing and deployment

We provide user-friendly UIs (user interfaces) that enable intuitive operation and visualization of the 3D data. This makes it easier for users with different levels of knowledge to analyze and interpret the data. Our APIs (Application Programming Interfaces) also enable seamless integration into existing environments.

Application in Multiview AI

The results of the depth calculation are a decisive factor for the further development and use of multiview AI. They are not only used for highly efficient data curation, but also form the basis for the training of multi-view AI models and the development of specialized AI models.

Application Examples

weldAI

Inspection of hairpin welding for electrical vehicles

pinOPTIX

3D connector inspection

PCB-IntelliScan

Recognizing and identifying components on printed circuit boards

honeycombINSPECT

Revolutionary inspection of honeycomb structures

lightfieldLAB

Light field laboratory system for the rapid generation of 3D point clouds

Advantages

Unsurpassed precision and detail

K|Lens Multiview depth calculation relies on sub-pixel precision, high-resolution depth maps and precise calibration to deliver accurate 3D data even under demanding conditions. This enables the capture of the finest details and detailed 3D reconstructions, which is particularly crucial in quality control.

Exceptional robustness and reliability

The K|Lens Multiview depth calculation offers high robustness against disturbances through a multiview approach and can be adapted to environmental conditions to deliver reliable results.

Maximum flexibility and adaptability

The K|Lens Multiview depth calculation is characterized by its exceptional flexibility and adaptability. A wide range of algorithms, from traditional to AI-based, enables the optimal selection for each specific task. The extensive range of parameters allows precise fine-tuning of the depth calculation to optimize accuracy, speed and robustness. In addition, custom adjustments can be made to object geometries and the software is flexible enough to be integrated into existing camera arrays.

High performance and efficiency for real-time applications

The K|Lens Multiview depth calculation impresses with its high performance based on optimized algorithms, accelerated processing by GPUs or edge devices and real-time capabilities, making it ideal for demanding applications such as robotics and in-line quality control.