Keywords: LiDAR; automotive industry; dust; rain; filtering algorithm; adaptability; LSLiDAR; point cloud; perception accuracy; operational efficiency; safety; breakthrough results; solid-state LiDAR; fiber LiDAR; corporate mission; intelligent driving; reliability; comprehensive solution; full-scene LiDAR; smarter machines; better life.
LiDAR requirements vary across different application scenarios, and this is especially true in the automotive industry. By utilizing LiDAR sensors installed on vehicles, a comprehensive understanding of the surrounding environment can be achieved. In this context, the adaptability of LiDAR systems to different environmental conditions is highly valued.
Mine Dust
Drive in the rain
Vehicles face challenges in maintaining stable and reliable performance in dusty or rainy weather conditions. For instance, driving on dusty roads or in mining areas where dust is kicked up, as well as driving in rainy or foggy conditions, can cause a decrease in the detection capability of LiDAR or result in unwanted interference points in the point cloud output for cars and robots. In such situations, the reliability of LiDAR systems faces an additional test.
In response to the challenges faced by vehicles and robots in operating environments with dust and rain, LSLiDAR has conducted extensive comparative experiments and calculations between dust and rain characteristics and point cloud algorithms. This effort has resulted in the development of a dust and rain filtering algorithm for LiDAR. By continuously optimizing and improving the algorithm based on real-world application requirements, LSLiDAR overcomes blind spots in algorithm and data learning, resolves issues related to the recognition of targets behind dust filters, and enhances operational efficiency, stable data output, and safety. Through these efforts, LSLiDAR has achieved breakthrough results.
This algorithm can be applied to any model of LiDAR from LSLiDAR. It has been tested and applied on the auto-grade hybrid solid-state LiDAR CH128X1 and the multi-line mechanical LiDAR C16/32, enabling them to possess dust, rain and fog identification filtering capabilities.
GIF 1 Dust Filter Features Off
GIF 2 Dust Filter Features On
As can be seen from GIF images 1-2, there is a lot of dust around or in front of LiDAR. However, the actual point cloud data of LiDAR shows that it can penetrate thick dust and detect targets behind it, even when the dust point cloud is significantly reduced.
GIF 3 Rainwater Filter Features Off
GIF 4 Rainwater Filter Features On
In images 3-4, during rainfall, scattered raindrops are treated as obstacles, hindering the path of vehicles or robots. The radar removes these raindrop points, relieving the data processing burden of the host computer. As such, CH128X1 provides more accurate perception and higher sensitivity.
LSLiDAR’s multi-line mechanical LiDAR C16/C32 has undergone a 4.0 upgrade, offering higher detection accuracy, more stable point clouds, and stronger light interference resistance. Additionally, it has been repeatedly tested for dust and fog, providing effective dust recognition and filtering, taking perception performance to a new level.
C16/C32 Dust Filter Features Off
C16/C32 Dust Filter Features On
C16/C32 Dust Filter Features Off
Additionally, this algorithm can also be applied to the Terminator Series 1550nm fiber laser auto-grade LiDAR and the CX Series 905nm auto-grade hybrid solid-state LiDAR. Ensures autonomous vehicles can safely operate in challenging environments such as rain, fog, and high dust levels. This aligns with LSLiDAR’s corporate mission of “Drive safer”.
LSLiDAR, as a global leading provider of full-scene LiDAR and overall solutions, adhering to the mission of “Drive safer, work smarter, and live better”, strive to empower autonomous vehicle with enhanced safety and reliability.
