CAJ | 학술논문

利用近红外光谱技术进行苹果霉心病的无损检测是目前的主流方法，其光谱与霉心病的病变程度密切相关。为了实现苹果霉心病病变程度的精确测量，设计并实现了一种基于计算机视觉的苹果霉心病病变程度测量方法。首先，基于Otsu 算法，利用果实区域与背景反差较大的特点，实现果实、病变与背景区域的分割，并对分割图像进行区域填充等预处理操作；然后，针对分割后包含病变的苹果图像，运用 Otsu 算法进行二次分割，分别提取腐烂病斑区域与整个苹果区域的面积；最后，计算出病变区域的比例，得到其病变程度。为了验证算法的有效性，利用多幅图像进行了测试，并与利用Photoshop 手工选取的实际病斑面积进行了对比。试验结果表明，利用该方法得到的腐烂程度误检率为8．87％，可以有效地实现苹果霉心病病变程度的测量。
이용근홍외광보기술진행평과매심병적무손검측시목전적주류방법，기광보여매심병적병변정도밀절상관。위료실현평과매심병병변정도적정학측량，설계병실현료일충기우계산궤시각적평과매심병병변정도측량방법。수선，기우Otsu 산법，이용과실구역여배경반차교대적특점，실현과실、병변여배경구역적분할，병대분할도상진행구역전충등예처리조작；연후，침대분할후포함병변적평과도상，운용 Otsu 산법진행이차분할，분별제취부란병반구역여정개평과구역적면적；최후，계산출병변구역적비례，득도기병변정도。위료험증산법적유효성，이용다폭도상진행료측시，병여이용Photoshop 수공선취적실제병반면적진행료대비。시험결과표명，이용해방법득도적부란정도오검솔위8．87％，가이유효지실현평과매심병병변정도적측량。
Near infrared spectroscopy technology is a common used nondestructive detection method of moldy core of ap -ples .This method is closely related to the extent of the decay area of apples .In order to realize precise measurement of apple mould core lesions degree of decay , the design and implementation of a measurement method of apple mould core lesion degree based on computer vision method was presented .Firstly , by using Otsu algorithm , fruit area and back-ground was segmented and the fruit region was filled and other pre-processing operation was used also .Then , the re-use of Otsu method was carried out according to characters of segmented apples .In order to verify the effectiveness of the al-gorithm, several sample images were tested and compared by calculating the detected areas with the Photoshop software detected areas , experimental results showed that the degree of decay false detection rate obtained by using the method presented in this paper was 8 .87%, which shows that it could realize the measurement of apple mould core effectively .