Render the closest to reality training images with our proprietary hyperspectral render.
Simulate faithfully the specific sensors that your system equips.
sensor decisions
When you train AI models with synthetic data, it is crucial to ensure they show a high degree of generalization to real-world data. Beyond a proper physics description of the synthetic 3D world, accurate modeling of the target sensor is essential to guarantee appropriate data matching in real operation. Anyverse’s sensor simulation uses an accurate hyperspectral light transport model. It combines a physics description of lights and materials in a 3D scene with a detailed simulation of sensor intrinsics. The sensor pipeline calculates energy’s propagation per wavelenth, considering many subtle physics effects such as the conversion of photons into voltage and RAW digital values before producing a final image.
low-pass filter, infrared cut-off filter, color filter array.
Exposure time, Conversion gain, Well capacity, Voltage swing, Dark Current and Noise (Shot Noise, Read Noise), Exposure (Global Shutter, Rolling Shutter), Motion Blur and Vibrations.
model provided by the user, or export the RAW data to an external ISP module.
using the ISET Toolbox from Stanford University.
Optical Systems:
Anyverse performs a geometric optical system simulation using ray tracing technology, capturing effects such as extreme optics distortion, lens shading, lens blurring (depth of field), complex assembly of multiple lenses, and more.
Multiple sensors:
Perception models are sensitive to different sensors. Anyverse™ technology allows you to simulate different sensors and ISP configurations to choose the right sensor for your perception system.
physically-based render engine
Anyverse™ hyperspectral render engine implements a pure spectral ray-tracing engine that computes the spectral radiance of every light beam interacting with materials in the scene, simulating lights, cameras and materials with physical accuracy. This allows for a very detailed simulation of the amount of light that is reaching the camera sensor to generate a final image containing all the spectral information from a 3D scene (using Anyverse’s sensor simulation pipeline).
Both, the hyperspectral render and the sensor simulation pipeline are the core of Anyverse’s synthetic data platform that helps our customers to generate the data they need to develop advanced sensing and perception systems.
Our render engine simulates energy as an electromagnetic wave all the way across the rendering pipeline to its final digital values. Light sources in Anyverse™ emit using a characteristic spectrum profile that depends on the type of light source (LED, incandescent bulb, sky, etc). The materials’ behavior is also wavelength-dependent, this allows for typical effects like dispersion. These features are also key and necessary for an accurate camera sensor simulation.
You have full control on several sensor parameters…
…To generate a raw image, and full control on Image Signal Processor (ISP) parameters.
The world is different depending on where you look, Anyverse™ has developed technology to recreate a diverse synthetic reality and to generate training and validation data. Generate thousands of variations of the 3D scene, change camera position, lighting, and weather conditions.
In addition, with Anyverse’s procedural engine and API, you can programmatically control your scenes and automatically generate all data variations.
