Machine Vision
CUSTOMER REQUIREMENTS

There are high manual screening 

costs and low efficiency.

Sorting cotton based on lightness is impossible for common color sorters.

Common color sorters have inaccurate identification of foreign fibers in cotton when relying solely on color.

Importing prism spectrographs from 

the current market incurs high costs.

The traditional algorithm has a detection speed of 10-15 m/s, resulting in low efficiency.

The detection rate is low, ranging from 70% to 80%.

SOLUTION

Detection and computation of foreign fibers in cotton using the MIIVII EVO Xavier.

Relying on the powerful encoding and decoding capabilities of the NVIDIA GPU, 

the EVO Xavier can analyze image data collected by the camera in real time, identify

foreign fibers and control high-flow nozzles to sort cotton based on detected lightness.

CONTACT US
FRAMEWORK
VALUE
01
This solution involves identifying foreign fibers such as mulch, nylon rope, woven bags, and hair using low-cost, high-definition cameras. An ultra-real-time computing network is implemented to improve the detection rate. Omission rate decreased from 5%-10% to less than 5%. Detection rate increased from 70-80% to greater than 90%.
02
Identifying foreign fibers such as mulch, nylon rope, woven bags, and hair using low-cost, high-definition cameras.
03
An ultra-real-time computing network is implemented to improve the detection rate.
04

Omission rate decreased from 5%-10% to less than 5%.
05
Detection rate increased from 70-80% to greater than 90%.