Machine vision
用户痛点
High manual screening costs
Low efficiency
Cotton sorting based on lightness, impossible for common color sorters
· Inaccurate identification of foreign fibers in cotton solely based on color
High costs in importing prism spectrographs from the current market
Detection speed 10-15m/s of the traditional algorithm, low efficiency
· 检Low detection rate, 70-80%
解决方案

Detection and computing of foreign fibers in cotton based on MIIVII EVO Xavier
Relying on powerful encoding and decoding of NVIDIA GPU, EVO Xavier can analyze the image data collected by the camera in real time, identify foreign fibers, and control the high-flow nozzles to sort cotton according to the detected lightness.

方案框架
方案价值
01
Creating a super-real-time computing network for 4K images depending on highly real-time performance and powerful computing, automatically synthesizing massive samples on the edge side and overcoming the difficulty that deep learning algorithms require mass data.
02
Identifying foreign fibers such as mulch, nylon rope, woven bags and hair using low-cost, high-definition cameras at low costs.
03
Ultra-real-time computing network to improve the detection rate
04

Omission ratio decreased from 5%-10% to < 5%
05
Detection rate increased from 70-80% to >90%