Vision-Guided Unordered Grasping of TV Wire Harnesses

Project Background

The home appliance manufacturing industry faces multiple challenges, including rising labor costs, increasing quality control pressures, and insufficient flexibility of production lines. To address these pain points, DexForce leverages embodied intelligence to empower the core links of home appliance production, and its spatial and embodied intelligence suite, powered by the DexVerse™ engine, acts as a catalyst to inject intelligent transformation into home appliance production lines.

Core Values

High-Precision Imaging

Enabled by the high-precision 3D camera that captures high-fidelity images of sheet metal parts.

Precise Positioning

Enabled by advanced vision algorithms that accurately locate deformed wire harnesses and guide robots for precise grasping.

Product Recommendations

Higher Precision 3D Imaging Capability

Based on self-developed high-precision 3D imaging algorithms and HDR technology, achieving Z-axis measurement accuracy up to 50µm at a working distance of 600mm, applicable to industrial inspection fields with high precision requirements

Integrated Imaging and Perception

Equipped with NVIDIA Jetson Nano core, built-in computing and control algorithms, endowing the camera with point cloud analysis capabilities and optimizing image quality; fast speed, more stable, enabling lightweight, edge deployment

Industrial-Grade Design, Addressing Real-World Scenario Challenges

IP65 dustproof and waterproof design; wide temperature design, supports operation in environments from -10°C to 50°C, adaptable to most industrial scenarios; passed certifications such as EU CE, RoHS, US FCC, etc., better coping with harsh industrial environments like humidity, electromagnetic interference, and dust

Algorithm Development for Extreme Conditions

Exposure modes specifically optimized for highly reflective object surfaces, effectively improving the camera's accuracy and stability in high-glare surface measurements. Proprietary compensation algorithms effectively reduce depth measurement errors, minimizing the impact of object surface textures on depth images

Vision-Guided Unordered Grasping of TV Wire Harnesses

Project Background

Project Challenges

Case Video

Core Values

Product Recommendations