Computer Vision based Automation

Computer Vision based Automation

Today, most industries, as well as businesses, are trying to take advantage of Computer Vision driven services to deliver superior customer experience, improved product quality, reduced cost, and increased security. Gadgeon over the years has been investing in fine-tuning our services and expertise to acquire, analyze, process, and understand digital images along with real-time data. This combined with robotic and embedded engineering expertise enables Gadgeon to automate manual labour intensive tasks efficiently.

Gadgeon works with clients in introducing computer vision-driven automation concepts. This involves modeling and mapping of current activities, tasks, risks, and challenges involved. This helps in identifying areas where efficiency can be improved, and business benefits are predicted.

Our robotic and computer vision-driven services are domain agnostic. It cognitively performs tasks similar to humans, but with better precision and zero errors. Our expertise includes dynamic controls of robots using real-time image processing involving deep learning and big data.

Our computer vision algorithms use Machine Learning or traditional vision techniques for performing the following tasks:

  • Measure and Analyze - Measure objects to determine accurate dimensions and analyze or compare them with reference data.
  • Pattern Recognition - Analyze patterns with reference data to determine GO or NO-GO. Example, analyze ‘Glue’ or adhesive patterns on a production line to determine the adhesive strength
  • Color-Based Analysis - Inspect object for color change in the visual or outside visual spectrum. For example, to check a fruit is ripe or not.
  • Object Detection for Pick and Place - Identifies a specific object in an image and giving the coordinates for dynamically controlling a Robot Arm in very high precision assemblies.
  • Inspection of Inaccessible Locations - Use rigid/flexible borescope/endoscope to view inaccessible areas not visible to the eyes and apply image processing. Specialized robotic arms and mechanical fixtures may be developed depending on a specific use case.
  • Dynamic Word, Icon or Shape Detection  - Detection of text/word or icons regardless of background or variations in positions. This could be used to test touch screen content using a robotic arm.
  • Image Classification - Groups images into different categories using big data-based ML or small data but with feedback learning.

Gadgeon has delivered computer vision driven automation and services for use cases from industrial domains like:

  • Manufacturing – Precision manufacturing, improved product quality, automated testing of assembled systems
  • Agriculture – Crop and plant monitoring at various stages, and segregation of produce
  • Automotive industry - Autonomous navigation systems and testing of infotainment systems
  • Healthcare – diagnostic and early disease detection like cancer
  • Security – IP camera integration and event detection

Gadgeon’s image processing along with machine learning provides benefits like:

  • Precise coordinates for Robotic control system to micro-meter level accuracy
  • Expertise with Borescope based image processing to allow robots to reach where cameras cannot reach
  • High product quality with minimal output variance.
  • Increased productivity from smart robots
  • Reduce cost and time through automation of labour intensive tasks
  • Special sensors are custom build that complements vision systems for specialized applications
  • Automated testing of HMI with complex content and physical knobs, buttons, etc.
  • Superior customer experience

View Detailed Case Studies
Let’s Start your IoT Journey
Contact Us