VisionFence is an advanced visual perception technology, designed to enhance the obstacle avoidance capabilities of robotic mowers. Utilizing cameras and AI algorithms, VisionFence enables robots to recognize and maneuver around obstacles like trees, flower beds, walls, and more with precision. This innovation represents a significant leap in the field of robotic mowing, solving the need for intelligent perception and transforming how devices interact with their environment.
The Journey of VisionFence: From Traditional Sensors to Vision-Based AI
In 2022, the lawn mowing industry primarily relied on ultrasonic and bumper sensors for obstacle detection, which had limitations in accuracy and couldn’t detect flat objects like picnic mats. VisionFence 1.0 marked the beginning of a new era in 2023 by combining camera technology and AI-powered obstacle detection. As the industry’s first visual-based AI system, it set a precedent for the integration of artificial intelligence in the lawn care market.
With ongoing advancements in AI and neural networks, the VisionFence team continues to lead the industry with innovative upgrades that align with the rising demand for smarter, autonomous mowing technology.
VisionFence in the H & i Series: Precision Obstacle Recognition and Boundary Control
VisionFence in the H and i Series brought a major leap in lawn care by introducing a fisheye single-camera AI system that delivers 150 degrees of vision. This cutting-edge system accurately identifies mowable areas, boundary edges, and common obstacles, setting a strong foundation for intelligent navigation. By combining real-time AI-powered recognition with visual data, VisionFence guides the mower around lawns with optimal efficiency and accuracy.
VisionFence in the X Series: Enhanced Detection with 300-Degree View and 3D Sensing
Building on its predecessor, VisionFence technology in the X Series expands its field of view to 300 degrees through a triple-camera setup integrated with ToF (Time-of-Flight) sensors. This upgrade not only broadens the mower’s vision but also leverages powerful processing capabilities to increase obstacle detection as well as avoidance precision. From trees and flower beds to irregular-shaped objects, VisionFence delivers elevated obstacle detection, allowing for smoother and safer lawn navigation.
A New Benchmark in Intelligent Mowing
Precision and Flexibility: A New Era in Obstacle Detection
While traditional sensors like ultrasonic and bumper systems often lack accuracy and flexibility, VisionFence leverages AI-driven visual recognition to deliver unmatched precision. By distinguishing between obstacles and accurately defining lawn boundaries, VisionFence ensures a flawless mowing area and adapts seamlessly to various landscapes.
Intelligent Obstacle Classification: Smart Adaptability for Any Scenarion
VisionFence technology goes beyond mere detection, employing advanced AI to classify obstacles by type—whether it’s a tree, flower bed, tools, suspended items, animals, or people. This capability allows the system to respond with customized avoidance strategies, enhancing both safety and mowing efficiency across diverse lawn environments.
Semantic Segmentation for Obstacle and Passable Area Detection
Using semantic segmentation algorithms, the lawn mower can distinguish between passable areas and obstacles, enabling basic obstacle avoidance for common items such as flower beds, trees, picnic mats, and more. This approach enhances the mower's perception by classifying multiple types of obstacles, making avoidance more flexible and dynamic.
Detection of Suspended Obstacles
In addition to the 2D segmentation limitations, the mower is equipped with a detection algorithm for suspended obstacles. This allows it to identify and avoid suspended obstacles like tables, chairs, and vehicles, preventing the mower from becoming stuck during its operation.
Human Detection with Adaptive Speed Control
The vehicle is equipped with a human detection algorithm that adjusts the mower's speed when people are detected in the VisionFence. The speed reduction varies based on the distance from the people, ensuring high-speed mowing while prioritizing people's safety.
Animal Friendly Mode
A specialized detection algorithm enables the mower to recognize animals within a 5-meter radius in front of the VisionFence. Upon detection, the mower automatically adjusts its mowing path, maintaining a minimum 1-meter safety distance between the mower and the animals during operation.
Keypoint Detection for Small Obstacles
We’ve developed a keypoint detection algorithm that targets small, narrow obstacles such as thin rods, pipes, or wires. This enables the mower to navigate around these obstacles with maximum mowing efficiency and avoidance.
Lens Cleanliness Detection
To ensure reliable image input for the perception algorithms, a dirt detection system is deployed on the VisionFence. It continuously monitors lens cleanliness, alerting the user when the lens becomes too dirty to perform accurate perception, ensuring optimal operation.
Share: