本文選題：疫苗毒株　切入點：成活性檢測　出處：《農(nóng)業(yè)機械學報》2017年10期

【摘要】：針對人工光照法在生物疫苗毒株(胚蛋)成活性檢測中,存在勞動強度大、效率低、準確性差的缺點,提出了一種仿生胚蛋成活性圖像無損檢測方法。針對以往研究中胚蛋圖像上下灰度不一、胚蛋蛋殼質(zhì)量不一等因素影響胚蛋血脈提取的問題,提出了一種基于最小類內(nèi)指數(shù)方差的自適應閾值圖像處理方法。通過對去噪后的胚蛋圖像進行圖像邊緣檢測及數(shù)學形態(tài)學處理,準確構(gòu)建了成活胚蛋主血脈二值形態(tài),通過計算胚蛋內(nèi)主血脈二值面積百分比判定胚蛋成活性。對胚蛋圖像進行識別實驗,結(jié)果表明,該系統(tǒng)識別一枚胚蛋用時0.093 s,胚蛋活性判定準確率為100%,可滿足滅活疫苗、凍干疫苗生產(chǎn)的活性準確率要求。
[Abstract]:The artificial illumination method in biological vaccine strain (Pei Dan) assay of activity, there are high labor intensity, low efficiency, poor accuracy, this paper presents a novel bionic embryo into active image lossless detection method. According to the previous research on gray image of eggs, egg shell quality and embryo factors affecting embryo egg blood extraction problem, we propose an adaptive threshold image processing method based on minimum class variance index. Based on the embryo egg image after denoising for image edge detection and mathematical morphology processing, accurate reconstruction of the survival embryo blood two value morphology, by calculating the embryo blood in two the value of the area percentage of eggs to determine activity. Experiments, the results showed that the embryo egg image recognition system, a 0.093 s embryo egg, embryo egg activity to determine the accuracy of 100%, which can meet the inactivated vaccine, freeze dried vaccine The accuracy of production is required.

【作者單位】： 中國石油大學(華東)機電工程學院;濱州學院電氣工程學院;
【基金】：國家自然科學基金項目(51405512) 山東省重點研發(fā)計劃項目(2016GNC112015)
【分類號】：TP391.41;TQ460.7
，

本文編號：1715261