Achieving full autonomy is a challenge the automotive industry has been grappling with for years, and the current biggest obstacle is developing a visual perception system that’s reliable in all weather and lighting conditions. The current most common solution that experts recommend is using a mix of sensors to achieve full autonomy, including cameras, LiDAR, radar, and others. But, each sensor still has its limitations. For example, LiDAR provides a 3D point cloud of surroundings but is limited in inclement weather, long-range detection, and cost. Meanwhile, cameras are limited in low light conditions, where shadows can obscure the image.
To address these limitations, Bright Way Vision developed GatedVision, a revolutionary technology that offers a more comprehensive approach in a single solution. GatedVision has the potential to bring us closer than ever before to full autonomy, so let's take a deep dive into how it compares to other technologies.
All vision systems rely on contrast to detect targets on the road, but shadows can obscure the image, making it difficult to differentiate between a dark spot caused by a shadow and a non-reflective material. This is especially important in autonomous vehicles, where the system needs to make split-second decisions based on the information it receives. Visual perception systems need to be able to differentiate between shadows and real obstacles to ensure the safety of passengers and any pedestrians or animals that may be on the road.
In order for machines or humans to detect an object, there must be a difference in brightness or color between the object and its surroundings, AKA contrast. This applies to all types of vision systems, including those that rely on artificial intelligence.
The challenge with shadows is that they can be mistaken for non-reflective materials or dark spots. This is where GatedVision comes in. GatedVision can modify the shadow in an image within a very short time period of less than 5 milliseconds. This means that the shadow can be spatially modulated or altered within a single frame of the image.
With GatedVision, it is possible to distinguish between a shadow and a non-reflective material or dark spot in the image within a single frame at a time. This is because the shadow will be modulated, while the non-reflective material or dark spot will not be affected.
Thermal:
Thermal cameras can detect heat signatures, but they are limited in detecting objects with similar temperatures to their surroundings. GatedVision, on the other hand, can detect small objects with low reflectivity by analyzing the shadows they cast.
LiDAR:
LiDAR is limited in detecting objects in inclement weather, where the laser beam can be scattered or absorbed by fog, rain, or snow. In contrast, GatedVision provides a clear image in all weather and lighting conditions, including darkness, rain, snow, and fog.
Camera:
Standard cameras are limited in low light conditions, where shadows can obscure the image. GatedVision, on the other hand, can differentiate between shadows and non-reflective materials, providing a clear image even in low light conditions.
Autonomous technology will remain at a standstill until a viable visual perception solution that can provide a clear image in all weather and lighting conditions is installed in all vehicles. GatedVision’s shadow detection capability is essential in autonomous vehicles, where split-second decisions are made based on the information received. GatedVision’s ability to differentiate between shadows and non-reflective materials provides a clear image even in low light conditions clearly outperforms all competitors. When compared to other technologies, such as thermal content, LiDAR, and cameras, GatedVision provides a more comprehensive solution that can detect small objects with low reflectivity, provide a clear image in all weather and lighting conditions, and differentiate between shadows and non-reflective materials.