Wayve’s Post

The recipe for realism in a 4D reconstruction model: 👱Low maintenance - PRISM-1 separates static and dynamic elements within a scene in a self-supervised way 🖼️Flexibility- PRISM-1 handles numerous agents and elements from pedestrians walking to vehicle doors opening 📈Scale- PRISM-1 remains efficient, even as scene complexity escalates. Read more about #PRISM1 here: https://lnkd.in/esybvhvm #EmbodiedAI #SelfDriving #AutonomousVehicles

In #3D reconstruction, moving vehicles can distort scene accuracy, creating redundant and #dynamic objects that impact model quality. Such interference can mislead subsequent data analysis and applications, making it vital to filter out these moving #vehicles. #Get3DMapper tackles this challenge by utilizing a method based on #multi-view consistency constraints to detect moving vehicles and their trajectories. Through #AI-driven algorithms, these dynamic objects are assessed and removed from the 3D model. The result? A cleaner, more accurate 3D scene model. By eliminating moving vehicles, Get3D Mapper ensures that 3D model data remains stable and precise, free from the influence of dynamic objects. This enhancement is crucial for various applications, including #autonomous driving navigation, #virtual reality environment creation, and urban #digital management. #3dmodeling #photogrammetry #computervision #mapping #innovation #3dreconstruction #AI #automaticdriving #digitaltwin

Unitree Robotics' H1 V3.0 Evolution sets a world record for the fastest full-sized humanoid robot, achieving a speed of 7.38 mph (3.3m/s). This surpasses Boston Dynamics' Atlas and showcases Unitree's commitment to innovation in legged robotics. The H1 V3.0 boasts dynamic movements, including fluid arm, leg, and torso motions, along with impressive choreographed routines. It stands at 71 inches tall, weighs around 100 pounds, and features a Unitree gear train, high-torque joint motor, depth camera, and 3D LiDAR sensor for vision. With the ability to leap to human height, navigate confined spaces, and exhibit whole-body coordination, the H1 demonstrates remarkable engineering prowess. Unitree aims to democratize legged robotics, with its Go2 quadruped starting at $1,600. Pre-orders for the H1 are ongoing, with deliveries set to begin by mid-2024 at a price range between $90,000 and $150,000. Positioned as a competitor to industry giants, Unitree's H1 offers an affordable yet advanced humanoid option #UnitreeRobotics #HumanoidRobot #H1Evolution #RoboticsInnovation #WorldRecord #LeggedRobot #BostonDynamics #AtlasRobot #AItechnology #RoboticEngineering

ultra realistic robotic horse with embedded horse characteristics to make it look real. https://lnkd.in/g-5-Xh_5 https://lnkd.in/g-5-Xh_5

Do you know the Mantis shrimp? This shrimp is a master of vision and we could learn from him. The Mantis shrimp possesses an extraordinary ability to perceive polarized light! It can discern transparent animals and fishes utilizing mirror camouflage, such as silverfish, which reflect the blue light of the ocean to render themselves virtually invisible. With an astonishing 16 color photoreceptors, some of which can discern polarized light, including circular polarization, the Mantis shrimp can not only detect the intensity of polarized light but also its direction. Remarkably, these creatures can even communicate secretly by emitting circularly polarized light visible only to themselves. In the image below, you can observe the stark contrast between the vision of an autonomous digital car camera, which relies solely on grayscale and poorly delineates object contours (left), and the use of camera simulating Mantis shrimp vision in the middle and right. The middle image depicts the intensity of light polarization, while the right image illustrates the angle of polarization. This simulation demonstrates how incorporating Mantis shrimp-like vision into autonomous vehicle systems can vastly improve their perception capabilities! While information derived from polarized light remains imperceptible to the human eye, it holds immense potential for enhancing various technologies. Advancements in Augmented Reality (AR) technologies may soon enable us to integrate this valuable information into our visual experiences. Say goodbye to face-planting into glass panes! Far from being confined to the realm of science fiction, these advancements are on the horizon, waiting to be harnessed. They have been present in nature for millions of years. #biomimetics #polarization #visionscience

Say hello to the R12 Rex CS. The world’s most powerful cleaning robot. Delving into the features of this world-class cleaning mechanism, it is designed for high performance and convenience. The LionsBot International R12 Rex CS cleans large areas while picking up debris with twin cylindrical brushes and side brushes. No more pre-sweeping needed. In large commercial spaces, this will lift debris on the floor, scrub the floor with 95kg pressure and dry within seconds. A 200m LIDAR range with 3D SLAM enables it to see further, map better and be more efficient. When it’s done, it auto-docks into its own refuel station — getting ready to go for the next round. Further marketing-leading features include: 🦾 Eliminate Hard Stains 🦾 Cleans Super Fast 🦾 Built To Last 🦾 All-Sensing 🦾 Multiple Control Options 🦾 Sustainable & Efficient Book in a demonstration at your premises today ➡️ https://lnkd.in/e_id8aTv #RexCS #LionsBot #RoboticCleaning #Cobotics #PristineGroup #CleaningSolutions #FutureTechnology #Sustainability

This train inspection robot is a great example of dull and dirty task. It uses LiDAR + 3D Camera vision system mounted on a robot motion platform. How does it work ? It uses a mechanical arm and vision system to inspect conditions like leakage, joint bars and rail ends and hidden anomalies at bottom of the train 🚆. What does this mean for the industry? 1️⃣ Efficiency: Boosting inspection processes without manual intervention. 2️⃣ Safety: Timely and comprehensive inspections mean identifying and eliminating hidden dangers in an operation. Developed by Shenhao Robotics International Video via Lukas M. Ziegler #RailInnovation #AutonomousRobot #TechAdvancements ******************************************** • 𝐅𝐨𝐥𝐥𝐨𝐰 𝐭𝐨 𝐬𝐭𝐚𝐲 𝐜𝐮𝐫𝐫𝐞𝐧𝐭 𝐨𝐧 𝐑𝐨𝐛𝐨𝐭𝐢𝐜𝐬, 𝐈𝐧𝐝𝐮𝐬𝐭𝐫𝐲 𝟒.𝟎 𝐚𝐧𝐝 𝐨𝐭𝐡𝐞𝐫 𝐓𝐞𝐜𝐡 𝐭𝐫𝐞𝐧𝐝𝐬

Hello LinkedIn community! 👋 I'm back with the second part of my autonomous robot simulation project using ROS1 Noetic! 🤖 📢 Quick Recap: In the first part, I showcased the initial steps of making my robot move in various ways, experimenting with hand gestures ✋, a PS4 joystick 🖲, and keyboard commands 💻. 🚀 Key Achievements in Part 2: *Sensor Fusion: Merging data from Lidar, Depth Camera, and IMU for comprehensive environment understanding. *Object Detection : Thanks to YOLO, the robot can now recognize and respond to objects in its surroundings. *3D Mapping : Utilizing depth camera pointcloud to create detailed 3D maps for autonomous navigation. *Arm Control: With MoveIt, the robot's arm moves effortlessly, responding to complex motion planning scenarios. 💁♂️ The robot can now move on its own, intelligently navigating through its surroundings, thanks to advanced sensors and sophisticated algorithms working together. 😊 A big thank you to the ROS community for sharing awesome posts! #ROSNoetic #RobotLearning #TechCommunity #RoboticSensors #SmartRobotics #ROSCommunity #Robotics #AutonomousNavigation #AI #RobotSimulation #MoveIt #Sensors #ROSDevelopers #Engineering #slam #computervision #3dmapping #depthcamera #IMU #rplidar #RobotControl #ProjectStart #TechInnovation #RoboticsFun #HandGestures #JoystickControl #KeyboardCommands #DIYRobotics #SimulationProject #opensource #opencv #gazebo #ps4 #navigation #autonomousdriving #selfdrivingcars

Vehicle Instance Segmentation with Object Tracking using Ultralytics YOLOv8 😍 🚀 🔥 🔗 Code: https://lnkd.in/dEvtNJEz ✅ Instance segmentation with object tracking assigns unique masks to individual tracks in an image, vital for precise object localization in tasks like object detection and autonomous systems. #computervision #segmentation #yolov8 #computervisionapplications

According to #ElonMusk, the quickest answer to driverless cars is camera vision instead of LiDARS because “humans drive with eyes & biological neural nets, [and making] sense that cameras & silicon neural nets are the only way to achieve a generalized solution to self-driving.”As technology advances, a Canadian company has built the first autofocus car camera that provides high-quality imaging across various temperatures. To learn more, head to the link below.