Reduced Production time by more than 80% for a LiDAR Manufacturer using Automation

The Need

A North American LIDAR manufacturing company was looking for a partner to automate the calibration process of dynamic calibration and alignment of an array of lasers to MICRON-level accuracy. The manual calibration and alignment process was taking over 24 hours per LIDAR slowing down production, and leading to very high production backlogs.

Our Solution

• A High-precision Image Vision system to assist in picking and placing by the Robotic Arm which is of the Denso VP series.
• Multiple cameras and a Borescope vision system were used to access areas inaccessible to regular machine vision cameras. 
• The computer vision guided the robotic arm to pick and place the PCB in the desired tray. The tray dimension was 10 X 10 CM, and the PCB dimension was 2 X 4 mm. 
• The robotic arm performed this task with high accuracy without damaging other PCBs of micrometer dimensions.
• Developed an image processing algorithm to predict the location of the PCBs with needed precision for the Robot to grab.
• Gadgeon developed an image vision system that controls a Robotic gripper to grab components and align them. 
• The Vision system was based on borescope imaging to reach inaccessible areas and a secondary camera to guide the first camera. 
• Device drivers for camera and image acquisition/manipulation form part of the complete automation software.
• Control of multiple cameras and lighting are implemented as embedded software.

Tools & Technologies

• Robot - Industrial grade Robot with 6 DOF
• Cameras and setup:  Blackfly S USB3: 20 MP, 18 FPS, Sony IMX183, Color
• uEye XS, USB 2.0, CMOS, 15.0 fps, 5.04 MPix, 1/4", Rolling Shutter
• Karl Storz rigid borescopes with 3.8 mm diameter
• Custom algorithms for distance measurement, calibration, and image processing
• Control of multiple cameras and lighting in embedded software

Images of Borescopes used