Scratch Detection on Speaker Surface

csdetb-icono

Customer

An inspection company that inspects various materials using High-end Technologies e.g. Computer Vision and Robotics.

Project Details

  • Duration: 1 Month

Technologies:

  • TensorFlow
  • Python
  • OpenCV
  • Industrial
  • GigE Cameras
  • Flask
problem-img

Problem

They faced multiple problems while finding defects on the Speaker Net Surface.

  • Missed Identification as it is Prone to Human errors
  • Low Accuracy using conventional AI models
solution-img

Solution

We have developed a custom AI Segmentation model that not only identifies/detects the scratches but also classifies them according to their nature.

  • High-end Industrial Camera captures the images and sends them to the Edge Processing Device for OCR.
  • This helps in prioritizing the seriousness of the defects and helps further to take decisive actions.

99%

The Model identifies the defects with 99% accuracy.

Insights on innovation

Stay updated with the trending and most impactful tech insights. Check out the expert analyses, real-world applications, and forward-thinking ideas that shape the future of AI Computer Vision and innovation.

March 6, 2026 - 5 minutes to read

The Real Cost of Blind Spots in Supply Chains (and How Visual Recordkeeping solves them)

At midnight, a high value shipment enters a regional logistics hub. By morning, the factory waited for those parts to halt production. The system shows the shipment as “received.” The yard team insists it was unloaded. The transporter says it left in perfect condition. No one is lying. No one can prove anything either. This […]

Read More

Ruchir Kakkad

CEO & Co-founder

March 6, 2026 - 6 minutes to read

How AI Vision is Transforming Operations across Industries

What ties these everyday moments together? It all looks connected… until you notice the pattern! Keep reading to know what? A truck entering a manufacturing campus. Crowds moving through a large exhibition hall. Containers rolling past a busy port gate. Players training across multiple tennis courts. Employees moving in and out of shared office spaces. […]

Read More

Ruchir Kakkad

CEO & Co-founder

January 27, 2026 - 7 minutes to read

The Hidden Value of Visual Recordkeeping in Logistics and Manufacturing

7:45AM at a busy automotive parts manufacturer. A delivery truck arrived at the dock. Usually, supervisors would take 10 minutes to scan paperwork. They’d also be keying in data. Instead, a camera quietly recorded the container ID, logged upon its arrival. And it tagged the shipment to a specific production line. All without a single […]

Read More

Ruchir Kakkad

CEO & Co-founder

Whatsapp Img