东北农业大学学报
東北農業大學學報
동북농업대학학보
JOURNAL OF NORTHEAST AGRICULTURAL UNIVERSITY
2009年
7期
106-110
,共5页
王立舒%陈月华%房俊龙%高汉峰
王立舒%陳月華%房俊龍%高漢峰
왕립서%진월화%방준룡%고한봉
模式识别%大豆%营养元素%智能诊断系统
模式識彆%大豆%營養元素%智能診斷繫統
모식식별%대두%영양원소%지능진단계통
pattem recognition%soybean%nutrition element%intelligence diagnosing system
营养元素是影响大豆品质和产量的重要因素,营养元素过量或缺乏均会时大豆产生严重影响.传统的营养元素诊断方法主要依靠人的主观经验,容易出现误诊、漏诊等情况,而传统的化学分析技术又会出现破坏植株、测试手段繁琐、周期长等缺点.文章通过分析大豆在营养元素过量或缺乏的不同情况下的植物特性,构建了基于模式识别的营养元素智能诊断系统,并提出了应用计算机视觉技术提取大豆形态特征,把拍摄的图像利用适当的方法进行图像分割、增强、平滑、滤波等处理,利用图像处理算法分割出叶脉、叶肉、叶缘.识别出颜色、纹理等形态特征,分析元素失衡时的颜色及纹理在叶片的不同部位表现,为今后进一步利用模式识别诊断大豆营养元素提供了发展方向.
營養元素是影響大豆品質和產量的重要因素,營養元素過量或缺乏均會時大豆產生嚴重影響.傳統的營養元素診斷方法主要依靠人的主觀經驗,容易齣現誤診、漏診等情況,而傳統的化學分析技術又會齣現破壞植株、測試手段繁瑣、週期長等缺點.文章通過分析大豆在營養元素過量或缺乏的不同情況下的植物特性,構建瞭基于模式識彆的營養元素智能診斷繫統,併提齣瞭應用計算機視覺技術提取大豆形態特徵,把拍攝的圖像利用適噹的方法進行圖像分割、增彊、平滑、濾波等處理,利用圖像處理算法分割齣葉脈、葉肉、葉緣.識彆齣顏色、紋理等形態特徵,分析元素失衡時的顏色及紋理在葉片的不同部位錶現,為今後進一步利用模式識彆診斷大豆營養元素提供瞭髮展方嚮.
영양원소시영향대두품질화산량적중요인소,영양원소과량혹결핍균회시대두산생엄중영향.전통적영양원소진단방법주요의고인적주관경험,용역출현오진、루진등정황,이전통적화학분석기술우회출현파배식주、측시수단번쇄、주기장등결점.문장통과분석대두재영양원소과량혹결핍적불동정황하적식물특성,구건료기우모식식별적영양원소지능진단계통,병제출료응용계산궤시각기술제취대두형태특정,파박섭적도상이용괄당적방법진행도상분할、증강、평활、려파등처리,이용도상처리산법분할출협맥、협육、협연.식별출안색、문리등형태특정,분석원소실형시적안색급문리재협편적불동부위표현,위금후진일보이용모식식별진단대두영양원소제공료발전방향.
Nutrition element is an important factor that affects quality and yield of soybean. Such as excessive or lacking nutrients will generate serious impacts on soybean. The traditional diagnostic methods of nutrients that mainly rely on subjective experience, so it is prone to causing misdiagnosis and missing diagnosis etc. The tradition chemical analysis technology may destroy individual plant, and its testing means is miscellaneous and trivial, which needs the long period and so on. The paper by analyzed the characters of the soybean nutrition elements form a nutrition element intelligence diagnosing system that based on the pattem recognition. According to soybean's structural difference in excessive or lacking nutrients state, and this paper also established nutrients intelligent diagnostic system which based on pattern recognition, and put forward the method to extract soybean structural features by a series of image processes (image segmentation, enhancing, smoothing and filtering etc.). The method could divide up leaf veins, mesophyll and leaf margin and identify the color, texture and many other structural features by image processing algorithms. By analyzing the elements in different conditions of color and texture in different parts of soybeans during the unbalance period, which provided the development direction of soybean nutrition element intelligence diagnosing system based on pattern recognition.