Vision-Guided Unordered Grasping for Nut Welding

Case Background

In automatic nut welding equipment, most of the nut loading, positioning, and welding processes are carried out manually—either by holding the nuts by hand or placing them into positioning fixtures manually. This results in low overall welding efficiency and high production costs. Leveraging the DexVerse™ engine, DexForce's spatial and embodied intelligence suite guides robots to orderly grasp and place four specifications of welded parts on the loading station, and accurately place workpieces at designated locations to complete the loading and unloading tasks.

Core Values

Sim2Real Model Training

The industry-exclusive Sim2Real technology enables vision models to be trained only by CAD models, disregarding of data collection or annotation, thereby enabling rapid and flexible production scaling.

Precise Positioning

Enabled by advanced vision algorithms that accurately locate welded parts and guide robots for precise handling.

Product Recommendations

Higher-Precision 3D Imaging Capability

Powered by in-house designed high-precision 3D imaging algorithms and HDR technology, the camera achieves a Z-axis measurement accuracy of up to 50μm at a working distance of 600mm. It meets the high standards of precision in industrial inspection applications.

Integrated Imaging and Perception

Equipped with the NVIDIA Jetson Nano module and built-in computing and control algorithms, the camera can perform point cloud analysis and optimize imaging quality. Its fast and stable operation enables lightweight and edge-side deployment.

Applicable to Challenging Industrial Scenes

With an IP65 rating for dust and water resistance and a wide operating temperature range of -10℃ to 50℃, the camera adapts to most industrial scenes. It is certified to CE, RoHS, and FCC standards, ensuring reliable operation in harsh industrial environments such as humidity, electromagnetic interference, and heavy dust.

Algorithm Development for Extreme Working Conditions

The dedicated exposure mode for highly reflective surfaces enhances the camera's measurement accuracy and stability. The proprietary compensation algorithm minimizes depth measurement errors and mitigates the effect of surface texture on depth imaging.

Vision-Guided Unordered Grasping for Nut Welding

Case Background

Case Challenges

Case Video

Core Values

Product Recommendations