Application

Let your robot understand what it sees.

With the support of artificial intelligence and deep learning, Mech-Mind Robotics offers cost-effective and easy-to-implement solutions for bin picking. Whether it's a classic reachpick out of a box with sorted parts, or random bin picking with chaotic and heterogeneous parts - Mech-Mind Robotics enables any type of machine tending controlled by machine vision. 

Our AI is specialized in detecting bulk material or partially structured workpieces and calculating the optimal gripping point. Taking the singularities of each robot into account, a collision-free motion sequence is automatically calculated. 

USP's

Core advantages of the Mech-Mind solution

Cost-effective 

In the field of AI-driven vision systems, made in China not only means state-of-the-art technology, but also an excellent price-performance ratio. Most of our 3D vision solutions provide our customers with an ROI of less than one year. 

Intelligent motion control

With optimal path planning and collision monitoring, Mech-Mind Robotics ensures particularly high stability and reliability even in challenging situations. Smart algorithms thus enable process-reliable solutions for every conceivable application in bin picking.

AI-supported and intelligent 

At the heart of Mech-Mind Robotics' all-in-one solutions is its sophisticated artificial intelligence. With the help of Deep Learning, robots recognize complex components and pick them up at the right place. This not only enables random bin picking of unsorted objects - it also  reduces the need for matching with CAD templates.

Quick and easy deployment

With the visualized and code-free programming interface Mech-Viz, robots can be controlled quickly and easily with just a few clicks. The entire motion sequence can also be simulated in advance and thus reliably checked. Mech-Mind Robotics thus meets the need for automation technologies that are as low-threshold and highly flexible as possible.

Process

Fast. Safe . Efficient.
More productivity in bin picking

If parts have to be picked from boxes in bulk or only partially pre-positioned, the use of bin picking is required. After the objects have been safely picked by a robot, they are precisely positioned for the next process step.

Precision

+/- 0.05 mm at 0.6 m distance, up to +/- 3.0 mm at 3 m distance.

Objects

Support of simple to complex geometries. Optimal recognition of all surface types, from matte to glossy, reflective or even partially transparent. Even colors do not play a role.

Cycle time

up to 3 seconds per cycle

Flexibility

Even dark or highly reflective parts with complex structures are reliably detected

Crankshafts are usually difficult to handle due to their structure and reflective surface. Thanks to the high-quality detection data of the Mech-Eye camera, the path planning can be easily adapted for the Mech-Viz robot controller. Collisions are thus reliably avoided.

Mech-Eye

Suitable cameras

Pro S

High speed and precision with fields of view up to 800 × 450 mm, suitable for KLT, cardboards and small boxes/crates.  Can reliably detect various objects (including objects with black or highly reflective surfaces). 

to our cameras

Pro M

High speed and precision with fieldsof view up to 1,500 × 890 mm, suitable for one Euro pallet or one mesh box. Can reliably detect various objects (including objects with black or highly reflective surfaces). 

to our cameras

Pro L

High speed and precision with fields of view up to 2.830 × 3.320 mm. Suitable for two Euro pallets or two mesh boxes.  Can reliably detect various objects (including objects with black or highly reflective surfaces). 

to our cameras

LSR L

High speed and high precision with fields of view up to 3.000 × 2.400 mm. Suitable for two Euro pallets or two mesh boxes, or for picking parts from a conveyor belt (conveyor picking).  Can reliably detect various objects (including objects with black or highly reflective surfaces). 

to our cameras

Successes

Case Studies

Construction equipment manufacturer: Bin Picking of track links

[Translate to Englisch:]

To increase efficiency with a guaranteed 24-hour production, Mech-Mind’s AI+3D industrial robot solutions enable the automation of bin picking of track links for a large construction equipment manufacturer. The 3D vision system recognizes and locates the track links piled in a deep bin, instructs the robot to grab one by one, double-check front and back side, and place them at a defined area in a fixed position. By using Mech-Eye Laser L, possible interferences of ambient light are significantly reduced and the camera can provide high precision pictures of a wide filed of view. Dozens of workstations have been automated, increasing the daily output up to over 1000 and reducing the costs by 60%.

Automotive industry: Machine tending of flange discs

[Translate to Englisch:]

Manual machine tending for heavy objects like flange discs is not friendly to the workers’ health and inefficient. To boost productivity, while maintaining production stability, a large flange disc manufacturer in China decided to automate its production line with Mech-Mind’s AI+3D industrial robotic solution. The superior 3D system guides robots to grab flange discs randomly piled in the deep bin and place them at their correct orientations and sides. This solution is suitable for picking flange discs of various specifications, can generate real-time data of the exact statuses and positions of each flange discs, and with the included intelligent motion planning system calculates the optimal path for the robots, so their movement is precise, stable, and collision-free.

Automotive industry: Bin Picking of inner gears

[Translate to Englisch:]

Challenging lighting and harsh working environment provide severe challenges for bin picking. For a large auto part factory Mech-Mind Robotics developed a 3D vision guided solution using Mech-Eye Laser L, which guides the robot to recognize and accurately grab inner gears randomly lying in the deep bin and place them at their designated positions. The solution can concurrently feed two machine tools and improved overall efficiency dramatically. The intelligent AI algorithms and the built-in motion planning and collision detection algorithms ensure an accurate and stable process.

Compatible robot brands (selection)