基于DSP机器视觉的缠辊检测系统

涤纶短纤后期处理中时常会发生缠辊。缠辊发生时，短纤后续表面会因缺丝而产生裂缝。辊筒缠丝越多，裂缝宽度就越大，辊筒就越容易损坏。针对如何实时、准确地检测缠辊这一问题，该文提出一种基于DSP机器视觉相结合的实时处理方案，介绍系统硬件组成和软件设计。在仿真实验中，系统对裂缝宽度以0.67 m/s增长的几幅仿真图进行检测，当发生报警时，检测时间为156 ms。

Wrap Detection System Based on DSP' Machine Vision

The fiber wrap always occurs in the later processing of staple fiber. When it happens, the surface of later fiber brings out some fissures because of lacking of silk. The more silk the roller wraps, the wider the fissure is, and the easier the roller damages. The paper presents a real-time solution based on DSP and machine vision for detecting the fiber wrap accurately in real-time, and also introduces the system hardware platform and software design. In the simulation test, some pictures whose fissure width increases at the speed of 0.67 m/s are detected, and it only takes 156 ms to find out the fiber wrap and send the alarm signal.