Geforce 256

Argos Solutions' plate scanners detect surface flaws, sort by quality, and aim to save time by optimizing image preprocessing on GPUs.

About us

We are a group of five enthusiastic data students who have received an exciting project from Argos Solutions AS regarding GPU-based image processing analysis.

Challenge

Argos Solutions provides scanners for plate manufacturers in the building materials and furniture industries. The function of these scanners is to detect defects on surfaces and sort according to quality criteria. The scanners incorporate line scanning cameras that deliver a continuous stream of images. Argos aims to leverage GPU in image processing to optimize and offload the workload on the CPU in their scanning processes. The increasing complexity associated with image processing has created a need for a more efficient solution. Argos acknowledges that traditional CPU approaches may be limiting and therefore views GPU as a potential solution to enhance processing efficiency. By exploring the transition from CPU to GPU, Argos seeks to address the technical challenges associated with handling image data and difference calculations.

Our Team

Sevag Hajji Narnian

Sevag Hajji Narnian

Software Engineer

Group Leader

Ali Hessen

Ali Hessen

Software Engineer

Documentation

Siem Ghebre Ghebrehiwet

Siem Ghebre Ghebrehiwet

Software Engineer

Product Owner

Iver Enget Nesbø

Iver Enget Nesbø

Software Engineer

Scrum Master

Abdiqani Abdullahi

Abdiqani Abdullahi

Software Engineer

Risk Analyst

Sprint 1

Sprint Image

We adopted scrum as our project management model, clarified project requirements, developed a risk template, and presented the project plan.

Sprint 2

Sprint Image

We started to gain understanding of UML, CUDA, and OpenCV, and successfully set up OpenCV and CUDA.

Sprint 3

Sprint Image

We began on developed algorithms for both CPU and GPU processing.

Sprint 4

Sprint Image

The team conducted Presentation 2, completed documentation, and enhanced project complexity.

Sprint 5

Sprint Image

Went on Easter vacation, completed the remaining tasks and began work on increasing complexity.

Sprint 6

Sprint Image

Began on the system for analysis and API, implemented blob detection, and started to performed post-processing.

Sprint 7

Sprint Image

Developed neural network for object detection as it detected, tracked and identified object, began on integration and a minimum viable product.

Sprint 8

Sprint Image

Finalize the product and documentation. Next week will include presentation, proofreading, clearing documentation and EXPO

Sprint 9

Sprint Image

Presentation, EXPO and Summer Vacation!!!!!!