食品科学 ›› 2004, Vol. 25 ›› Issue (10): 95-100.

• 基础研究 • 上一篇    下一篇

基于计算机视觉的肉色量化研究(英文)

 赵红霞, 周丰昆, 刘静波, 孙佳石   

  1. 中国科学院长春光学精密机械与物理研究所,吉林大学
  • 出版日期:2004-10-15 发布日期:2011-10-24

Study on the Color Quantification of Pork with Computer Vision System

 ZHAO  Hong-Xia, ZHOU  Feng-Kun, LIU  Jing-Bo, SUN  Jia-Shi   

  1. 1.Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences;2.Jilin University
  • Online:2004-10-15 Published:2011-10-24

摘要: 猪肉的颜色是判断其食用安全性非常重要的指标。肉的变质程度与颜色的变化呈现密切的关系。目前,肉色的分级判断通常是在自然光下,依据标准肉色等级图,靠人眼观察进行对比感官评定,这种主观评价方法存在着一定的人为误差,难免使判定结果不够客观和准确。因此,采用先进的方法、技术和仪器,通过对食用肉的颜色和颜色变化进行定量化测定和计算,可以获得客观评价。依据颜色理论,建立了一套用于猪肉颜色量化的计算机视觉系统,光源选择CIE推荐的标准照明体A,该系统能对猪肉的颜色、色差进行采集、处理、显示和评价,此方法具有操作简单、无损伤、测试速度快和费用低等特点,摆脱了颜色检测的主观性和模糊性,实现了蔬菜颜色评定和控制的客观化和定量化。

关键词: 猪肉, 色差, 计算机颜色量化系统, 神经网络

Abstract: The color of pork is a very important index on judging its edible safety. The degree of bad meat and the color changepresent close relation. Now the color of pork is usually judged according to the standard color card of pork relying on person eyesto observe under natural light. This kind of subjective assessment technique has certain artificial error leading to inaccurate result.Therefore, the objective appraisement should be obtained by the quantitative measuration and calculation at the color and thecolor change of edible meat with advanced method, technology and instrument. According to the color theory, a computer visionsystem used in porky color quantification measurement was established, the standard illumination body A that is recommendedby CIE was selected as the light source. This system can carry through collection, management, display, and appraisement ofporky color. This method takes on the characteristic of handling simple, nondestructive wound, fast measurement speed and lowcost etc, gets rid of the subjectivity and illegibility during color measurement, and realizes the impersonality and quantitativeappraisement of porky color.

Key words: pork, color difference, computer visual system, neural network