首页> 外文会议>IEEE International Conference on Granular Computing >An efficient detection method for rare colored capsule based on RGB and HSV color space
【24h】

An efficient detection method for rare colored capsule based on RGB and HSV color space

机译:基于RGB和HSV色彩空间的稀有色胶囊有效检测方法

获取原文

摘要

Rare colored capsule (RCC) is the capsule which is mixed in the normal product but with different color. To detect this kind of capsules online, a method based on RGB color space and using HSV color difference formula is proposed in this paper. Considering HSV color space with more balanced color-difference perception, the original RGB formatted capsule image is transformed to HSV color space for accurate detection. To guarantee the on-line detection efficiency, the color space transformation and the color difference calculation are not on pixel level. A set of R, G, B component values are acquired to represent the color of the capsule by a novel average gray calculation method based on histogram. Thus the RGB to HSV transformation is only need to process the set of R, G and B component values. The transformed H, S and V component values are then applied to calculate the color difference. By comparing the color difference with the given threshold, the final detection result for RCC is achieved. Compared with the usual color recognition method based on HSV, which transform color space in the whole image, the presented method is time saving. Online test results show that the proposed method has high accuracy and efficiency.
机译:稀有色胶囊(RCC)是混合在正常产品中但颜色不同的胶囊。为了在线检测这种胶囊,提出了一种基于RGB色彩空间并利用HSV色差公式的方法。考虑到具有更平衡的色差感知的HSV色彩空间,将原始RGB格式的胶囊图像转换为HSV色彩空间以进行准确检测。为了保证在线检测效率,色彩空间变换和色差计算不在像素级。通过基于直方图的新型平均灰度计算方法,获取一组R,G,B成分值以表示胶囊的颜色。因此,从RGB到HSV的转换仅需要处理R,G和B分量值的集合。然后,将转换后的H,S和V分量值应用于计算色差。通过将色差与给定的阈值进行比较,可以获得RCC的最终检测结果。与通常的基于HSV的颜色识别方法相比,该方法可以在整个图像中转换颜色空间,与传统的基于HSV的颜色识别方法相比,该方法节省了时间。在线测试结果表明,该方法具有较高的准确性和效率。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号