Sensor modelling for an AUV shows both RGB and LiDAR sensor fusion.
A dusty RGB sensor is simulated here with direct sunlight and deep shadows affecting the imagery. Using SynCity, you can employ sensor conditions including dust, water, and fog to ensure your system performs correctly under heavy weather conditions and edge cases.
CVEDIA creates realistic day/night cycles in SynCity and synthetic datasets.
A 360 degree camera is seen in an ADAS environment here, mimicking a highly contrasted RGB sensor.
A drone navigating through a forest is here built in SynCity using RGB and LiDAR modelling. Sensor fusion allows the simulations to perform together in real-time.
A weather edge case is seen here with large water drops detracting from an RGB sensor. Deploying autonomous systems can be made safer with edge case simulations like these.
Realistic waterdrops and lightning are seen in a simulated RGB sensor here. At CVEDIA we know that validation lies in tiny details like water refraction effects, and we put a lot of effort into recreating them.
A demo of a segmentation map synthesized in SynCity.
LiDAR is synthesized for ADAS here in a simulated San Francisco street.
An aerial perspective of a savannah environment is simulated here using NIR sensor modelling.
An infrared simulation of a night time ADAS environment.
LiDAR point clouds are simulated here alongside semantic segmentation in a forested ADAS environment.
Here we play with a simulated maritime environment using the SynCity interface. Weather patterns including fog, rain, clouds, and sun are adjustable with pull bars. Sensor conditions including water drops and sun glare are also able to be tested.
This synthetic ADAS dataset shows a high level of entropy including car models, humans, and city infrastructure. The light here is designed to mimic what a sensor would perceive – there’s a relative dimming when under shade, and a reflection of sunlight in the pavement.
In this simulation, we synthesize a drone flying over a city intersection. Vehicles are auto-annotated with 3D bounding boxes.
Five sensors combine for effective sensor fusion in this example of synthetic data for the security and defence industry, shown in the SynCity platform.
LiDAR simulation, shown here, is able to effectively sense surrounding vessels.
SynCity integrates with third party software – here, vehicle, electrical pole, and building segmentation in an ADAS simulation is integrated with ANVEL.
Several radar demos are synthesized next to an RGB parking lot simulation.
Body part segmentation, thermal imaging, and people counting simulation is seen in this combined retail environment.
A thermal sensor simulation is seen here on a synthetic highway.
CVEDIA recreates realistic day/night cycles for the creation of datasets that include all lighting conditions.
An RGB night search and rescue scenario is simulated here including camera zoom and water drops on the sensor.
Light reflection from car headlights are seen in this RGB night simulation.
Here, a flyover of an urban environment shows segmented buildings, trees, cars, and green space. On the left, a realistic day/night cycle is followed and shadows and lightning is employed.
A quick introduction to SynCity.
Here a car chase seen from a helicopter is simulated with both RGB and thermal sensors. CVEDIA’s partnership with FLIR Systems, the world’s leading thermal sensor producer, allows us to better predict and synthesize thermal sensor behaviour.