High-end quality control for the agricultural sector

Automated detection via AI-driven gigapixel analysis

The ‘mite camera’ brings agricultural inspection to a higher level with AI-driven gigapixel analysis. A solution was developed with NiftyBits that combines fine mechanics with machine learning to make details that are invisible to the eye measurable. This resulted in quick and objective inspections with total precision control.

Facts & Results

Gigapixel-reconstruction: From automated stitching of hundreds of macro photos to one extremely detailed full picture.
Automated Detection: AI-driven recognition for autonomous counting of micro-objects or identification of visual indicators.
Full-Stack development: Realization of both the complex mechanics and the powerful desktop application and control library.

Client

Technology

Computer Vision

Machine Learning

Precision Engineering

Gigapixel Imaging

Agricultural Innovation

Innovative quality control for the agricultural sector

Precision is essential for quality control in the agricultural industry. The challenge was to development an inspection system that could operate at a microscopic level while simultaneously offering an overview of the whole product. Therefore, NiftyBits developed the “mite camera”: a complete solution that combines hardware, fine mechanics, and advanced software to make invisible details measurable.

Extreme precision: The technique behind the image

The mite camera has the ability to capture micro level details in a composite gigapixel image. The system rapidly captures a vast number of zoomed-in photos which are automatically and seamlessly stitched together using advanced algorithms. An image is created through this process with a resolution that is impossible to create with standard cameras. To guarantee such optical precision, NiftyBits has designed and built the complete housing unit and all of the technical moving parts. The high-quality mechanics ensure that the camera moves with precision over the object to be captured.

Machine Learning & Analysis

The system is trained by machine learning to automatically recognize patterns, find defects, or count small objects (like mites) in the gigapixel photos. This replaces time-consuming tasks done by hand with an objective and rapid digital analysis. The administrator operates the system through a custom-made desktop application that translates the complex computing power into an intuitive user interface.

Integration and Customization (API)

Along with the physical device, NiftyBits has also developed a custom software library and API. With this software, the users have full control over the hardware and they can directly integrate the camera functions into their own business processes and software infrastructure. From the physical screws in the housing unit to the complex AI algorithms in the cloud, the mite camera proves how custom machine building and modern software engineering push the boundaries of inspection technologies.