English English German German
Computer Vision and Machine Learning

Edge compute has become more powerful and cost effective. Dedicated hardware from Intel, Google, Nvidia enable righ ML inference logic to be run on the edge device leading to effective Vision based application to improve efficiency, safety and performance in Industrial process

Assembly Line Process Adherance
  • Assembly lines have manual steps where quality control is challenging – for example an automated door assembly. 
  • A Visual inspection is not always possible due to concealed parts. Also it is not a scalable solution
  • A CV based approach monitors the process through a camera and detects hand positions and patterns. 
  • A ML process then maps these patterns to a known-pattern and scores the process adherence. 
Quality inspection. In-line & real-time 
  • A Camera captures the image of parts passing over the assembly line. 
  • This image is processed by a ML inference engine running within the Edge device to detect anomalies and defects
  • The ML model  itself is pre-trained with sample images classifying good parts and defects. 
  • The Edge device connects to the robotic-ARM over a Siemens S7 protocol to instruct segregation of good and defective parts
Safety & Health
  • Surveillance camera feeds from factory floor are fed to an edge device running ML inference
  • The edge device determines the currently safe areas and persons in unsafe areas to create audible or visible alerts
  • Use ML to determine usage of safety helmets and other safety gears during critical operations
  • Improve safety, health and working conditions of factory staff and reduce risk