纺织学报 ›› 2018, Vol. 39 ›› Issue (06): 131-135.doi: 10.13475/j.fzxb.20170803105

• 管理与信息化 • 上一篇    下一篇

基于深度学习的籽棉中异性纤维检测方法

  

  • 收稿日期:2017-08-18 修回日期:2018-03-07 出版日期:2018-06-15 发布日期:2018-06-15

Detecting method of foreign fibers in seed cotton based on deep-learning

  • Received:2017-08-18 Revised:2018-03-07 Online:2018-06-15 Published:2018-06-15

摘要:

针对籽棉图像阴影多、常规图像处理方法难于识别的问题,以去除棉叶、棉壳等有机杂物的籽棉为样本,将不同颜色、形状、尺寸的12种常见异性纤维和籽棉样本随机地分布在运转中的传送带上,采用线扫描相机获得发光二极管(LED)照明的籽棉图像520张,“LED+线激光”双光源照明的籽棉图像1 148张。然后采用一种由13个卷积层、13个采样层和4个池化层构成的Faster RCNN深度学习人工神经网络,对 2 种成像方法获得的籽棉图像进行基于人工智能的网络训练,再进行异性纤维检测验证。实验数据表明,LED照明和“LED+线激光”双光源照明条件下,籽棉图像中的异性纤维的检出率分别达到了90.3%和86.7%,特别是LED照明条件下对白色异性纤维进行识别,其识别率由5.9%提升到了90.3%。

关键词: 籽棉, 异性纤维, 深度学习, 人工智能, 图像处理

Abstract:

In order to detect foreign fibers in seed cotton with heavy shadows, seed cotton samples without cotton shell and leafs, and 12 types of foreign fibers with different colors, shapes and sizes were randomly distributed on a moving convey surface. And then, 520 seed cotton images were obtained under the illuminations of light emitting diode (LED) and 1 148 images were obtained under the illuminations of double light source of LED +  linear laser by a color line-scan camera. Then Faster RCNN deep-learning neural networks composed of 13 convolutional layers, 13 sampling layers and 4 pooling layers were constructed. After training, the neural networks were used for detecting foreign fibers in the two types of seed cotton images respectively. The experimental results indicated that the detecting rates of the targets in the images under the illumination of LED and LED + linear laser are 90.3% and 86.7%, respectively, by the Faster RCNN. Especially, the detecting rate of white color foreign fibers increase from 5.9% to 90.3%.

Key words: seed cotton, foreign fiber, deep-learning, artificial intelligence, image processing

[1] 汪珊娜 张华熊 康锋. 基于卷积神经网络的领带花型情感分类[J]. 纺织学报, 2018, 39(08): 117-123.
[2] 陆奕辰 王蕾 唐千惠 潘如如 高卫东. 应用图像处理的纱线黑板毛羽量检测与评价[J]. 纺织学报, 2018, 39(08): 144-149.
[3] 王雯雯 高畅 刘基宏. 应用卷积神经网络的细纱断纱锭位识别[J]. 纺织学报, 2018, 39(06): 136-141.
[4] 王雯雯 刘基宏. 应用优化霍夫变换的细纱断头检测[J]. 纺织学报, 2018, 39(04): 36-41.
[5] 王传桐 胡峰 徐启永 吴雨川 余联庆. 改进频率调谐显著算法在疵点辨识中的应用[J]. 纺织学报, 2018, 39(03): 154-160.
[6] 牟新刚 蔡逸超 周晓 陈国良. 基于机器视觉的筒子纱缺陷在线检测系统[J]. 纺织学报, 2018, 39(01): 139-145.
[7] 王晓予 向军 潘如如 梁惠娥 高卫东. 服饰刺绣图案的自动提取与色块分割[J]. 纺织学报, 2017, 38(09): 120-126.
[8] 路凯 钟跃崎 朱俊平 柴新玉. 基于视觉词袋模型的羊绒与羊毛快速鉴别方法[J]. 纺织学报, 2017, 38(07): 130-134.
[9] 张继东 薛元 张杰 郭明瑞 魏晓婷 高卫东 . 应用混色纱纹理信息的纬编针织物模拟[J]. 纺织学报, 2017, 38(07): 148-154.
[10] 张宁 李忠健 潘如如 高卫东 韩要宾. 采用色纺纱图像的真实感色织物模拟[J]. 纺织学报, 2017, 38(05): 37-42.
[11] 韦平 张林 刘翔 王冬 苏真伟. 籽棉中异性纤维的双光源成像检测方法[J]. 纺织学报, 2017, 38(04): 32-38.
[12] 刘成霞 韩永华. 模拟实际着装的织物抗皱性测试方法[J]. 纺织学报, 2017, 38(03): 56-60.
[13] 王传桐 胡峰 徐启永 吴雨川 余联庆. 采用Gabor滤波簇和等距映射算法的织物疵点检测方法[J]. 纺织学报, 2017, 38(03): 162-167.
[14] 李冠志 赵强 汪军 Gong Hugh . 纱线截面压缩变形仿真与验证[J]. 纺织学报, 2017, 38(02): 184-190.
[15] 李忠健 潘如如 高卫东. 应用纱线序列图像的电子织物构建[J]. 纺织学报, 2016, 37(3): 35-0.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!