Overview
The RealMan Document Center provides the download of five SDK descriptions, deployment, sample Demo, and secondary development resources regarding the RM robotic arm combined with AI, including multi-modal recognition, item segmentation, item pose, item tracking, and vision-based vertical robotic grab. It enables developers to fully understand the combination of the RM robotic arm with AI, lowering the development threshold and improving the development efficiency.
Target user
- Integration developer or robotic arm developer: Quickly apply the robotic arm with a vision to the assembly line, warehousing and logistics, book management, electronics assembly, medical device operation, etc. to improve the development efficiency in automation projects and reduce the development cycle.
- Development engineer for image recognition: Develop innovative projects and products using visual identity to increase the intelligence and automation of products.
- Researcher: Expand the development of algorithms based on vision for large model applications, etc.
- Educator: Design experimental projects for robotics, robot operating systems, machine vision, AI, and other courses for students to apply theoretical knowledge to practice.
Main content
SDK Developer guide
Multi-modal recognition: Refers to an RGBD fusion target detection method based on depth compensation, which mainly uses RGB and depth dual-modal information recognition for item identification and segmentation.
Item segmentation: Refers to the segmentation of any item in the screen to extract its outline information. With SAM encapsulation and specific adaptation, it generates the segmentation Mask without training and quickly grabs any item with the pose generation algorithm.
Item pose: Refers to the inference of the position and orientation of any item in the three-dimensional space based on its two-dimensional image and depth information. It has important practical application value in the fields of automated retail and item grab.
Item tracking: Refers the tracking using the long-term video object segmentation based on the Atkinson-Shifflin memory model, with automatic labeling and efficient segmentation functions, which is widely used in intelligent video processing in many fields. It is commonly used to track an item in continuous video frames or real-time images.
Vision-based vertical robotic grab: Refers to the recognition and grasp of items on the desktop by RM robotic arms based on vision, which can grab items less than or equal to 1 cm height on the desktop.
Visual servo: Refers to the visual servo realized by RM robotic arms based on vision. The visual servo controls the robotic arm through visual feedback for precise localization and automatic operation in dynamic environments.
Resource access
The SDK source code download and training weight file is provided for users to carry out agile deployment.
Contact information and support
- For questions about RM robotic arm and AI applications, please send an e-mail to sales@realman-robot.com.
- For the latest information on RM robotic arm and AI applications, please follow our Wechat official account.
