高粱圖像的分形特性研究


頁數(shù)：20 字數(shù)：11579


目 錄
摘要 1
關(guān)鍵詞 1
Abstractt. 1
Key words 1
0引言 1
1材料與方法 5
1.1樣本采集 5
1.2研究方法 5
1.2.1分維數(shù)的計算 5
1.2.2千粒重的測定 6
2結(jié)果與分析 7
2.1 背景分割 7
2.2計算分形維數(shù) 10
2.3不同品種高粱的分形維數(shù) 13
2.4高粱籽粒像素值與千粒重的關(guān)聯(lián) 14
3討論 16
3.1不同品種分維數(shù)比較 16
3.2高粱籽粒像素與千粒重的關(guān)系 16
4結(jié)論 17
參考文獻 18



摘要：為高粱品種分類，基于分形理論，對4種高粱籽粒圖像進行分形分析，比較其分形特征，并將高粱千粒重和籽粒圖像像素進行關(guān)聯(lián)?；诨叶戎狈綀D的閾值進行背景分割；基于分形原理計算分形維數(shù)，得到紅高粱、兩糯一號、遼粘3號、背面分形維數(shù)平均值分別為2.5257、2.6185、2.6344、2.6295，腹面分形維數(shù)平均值為2.5359、2.6344、2.6252、2.6309；除外兩糯一號，這些高粱品種的腹面分維數(shù)都比其的背面分維數(shù)要大，遼粘3號與甜高粱分形維數(shù)接近，但兩者背、腹面分形維數(shù)有差別；籽粒像素與千粒重呈現(xiàn)出同樣的增減性。結(jié)果表明，分形特征和籽粒像素與千粒重的對應(yīng)關(guān)系可以為高粱品種分類以及品質(zhì)區(qū)分提供依據(jù)。

關(guān)鍵詞：高粱；分形；分形維數(shù)；像素；千粒重

Fractal characteristic of 4 kinds of representative sorghum
Student majoring in agricultural electrification and automation
Tutor
Abstract: Using fractal method, the fractal dimensions of 4 different kinds of representative sorghums images are calculated to compare their fractal characteristics. Additionally, One-thousand-grain weights of are measured to assess the relationship between the pixel and the thousand-grain weight. It may provide references for sorghum assortment. The background segmentation is based on gray threshold; The rear fractal dimensions average values are 2.5257、2.6185、2.6344、2.6295 and values are 2.5359、2.6344、2.6252、2.6309, The ventral fractal dimensions are larger than the ventral except for Liangnuo-1.The pixel and the thousand-grain weight are monotonic. The result indicated that it might provide references for sorghum assortment and quality through the fractal characteristics and the relationship between the pixel and the thousand-grain weight.

Key words: sorghum; fractal; fractal dimension; pixel; thousand-grain weight