亚洲一区欧美在线,日韩欧美视频免费观看,色戒的三场床戏分别是在几段,欧美日韩国产在线人成

基于機(jī)器視覺的玉米果穗性狀參數(shù)測量方法研究
作者:
作者單位:

作者簡介:

通訊作者:

中圖分類號:

基金項目:

北京市重點實驗室2018年度科技創(chuàng)新基地培育與發(fā)展專項


Measurement Method of Maize Ear Characters Based on Machine Vision
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 圖/表
  • |
  • 訪問統(tǒng)計
  • |
  • 參考文獻(xiàn)
  • |
  • 相似文獻(xiàn)
  • |
  • 引證文獻(xiàn)
  • |
  • 資源附件
  • |
  • 文章評論
    摘要:

    在玉米育種、田間測產(chǎn)和提高玉米產(chǎn)量的過程中,均需要對玉米果穗考種,即需要對玉米果穗的穗長、穗粗、穗行數(shù)、行粒數(shù)和穗粒數(shù)等性狀參數(shù)進(jìn)行測量。人工考種不僅花費大量的人力物力,而且在考種過程中普遍存在人工勞動強度大、觀測效率低、人為干擾導(dǎo)致測試結(jié)果不客觀及不準(zhǔn)確等問題,在很大程度上限制了考種的速度與精度。針對上述問題,利用所研制的自動考種設(shè)備和機(jī)器視覺方法,通過USB工業(yè)相機(jī)獲取玉米果穗單面性狀彩色圖像,利用|B-R|模型、(G+B)/2模型將彩色圖像分別進(jìn)行灰度化,利用改進(jìn)后的一維最大熵閾值分割方法對灰度圖像進(jìn)行二值化,分別得到果穗輪廓二值圖像和果穗特征二值圖像;通過輪廓二值圖像計算果穗放置后的傾斜角,實現(xiàn)果穗輪廓二值圖像和特征二值圖像的自動糾偏;通過相機(jī)標(biāo)定,得到單位像素對應(yīng)的實際值,進(jìn)而得到穗長及穗粗;通過提取局部籽粒特征二值圖像,利用水平黑背景點掃描及對掃描曲線的修正獲取穗行寬度,通過穗行數(shù)修正模型得到果穗的穗行數(shù);通過提取局部單行籽粒特征二值圖像,利用垂直黑背景點掃描及對掃描曲線的修正得到行粒數(shù);根據(jù)行粒數(shù)和穗行數(shù)得到穗粒數(shù)。試驗結(jié)果表明,穗長和穗粗平均測量精度分別為98.05%和97.99%,穗行數(shù)測量正確率為95%,行粒數(shù)平均測量精度為96.29%,穗粒數(shù)平均測量精度為95.67%,和實際值相比,穗粗、穗長、行粒數(shù)及穗粒數(shù)的測量值差異無顯著性。單穗玉米果穗機(jī)器視覺平均測量速度為600ms/穗,考種設(shè)備測量速度為6s/穗,能夠滿足自動考種設(shè)備的使用需求。

    Abstract:

    In the process of maize breeding, yield tests and the improvement of maize production and the examination, including the measurement of the length, diameters, row numbers, grain numbers per row and grain numbers of ears, is necessary. However, manual examination utilized for a long time not only needs to spend a lot of manpower and resources, but also has many problems, such as high labor intensity, low efficiency of observation and the nonobjective and inaccurate results caused by human interference, greatly limiting the speed and accuracy of the operation. Therefore, an automatic equipment of examination with the machinevision was presented. The colour images of singleface characters of maize ears were obtained by industrial cameras through USB. Then the model |B-R| and (G+B)/2 were respectively applied to gray the colour images. After that, the method of segmentation of onedimension maximum entropy was used to achieve binaryzation, obtaining binary images of contours and features of ears separately. Moreover, these two kinds of images were corrected automatically by the calculation of angles of ear contours of binary images. Based on the calibration of cameras, the unit pixel corresponding to the actual value could be gained and then the length and diameters of ears could be calculated. In addition, the width of rows of ears could be got by the scanning of points of horizontal black background and correction of scanning curves, according to the extraction of local features of binaryzation images. The number of rows of ears could be obtained by the modified model of numbers of rows. Furthermore, by the extraction of binaryzation images of local features of single line of grains, the number of grains of rows could be obtained, based on the scanning of the points of black background and its modified curve. Finally, the total number of grains of ears could be computed by the numbers of rows and that of grains in single row. The experimental results showed that the average accuracy of measurement of ear length and ear diameter were 98.05% and 97.99%; the correct rate of measurement of row numbers was 95%; the average measurement accuracy of the number of grains per row was 96.29%; and the average accuracy of measurement of grain numbers was 95.67%. Furthermore, Ttest was conducted to compare the difference with the standard value, demonstrating that there was no significant difference and the equipment was of reliability. The average speed of measurement of the whole ear was less than 600ms per ear, and the measurement speed of the test system was within 6s per ear, meeting the requirement of the automatic equipment of examination. This research provided the basis of equipment and technology for the modern seed industry, even for the development of agricultural information technology.

    參考文獻(xiàn)
    相似文獻(xiàn)
    引證文獻(xiàn)
引用本文

吳剛,吳云帆,陳度,李寶勝,鄭永軍.基于機(jī)器視覺的玉米果穗性狀參數(shù)測量方法研究[J].農(nóng)業(yè)機(jī)械學(xué)報,2020,51(s2):357-365. WU Gang, WU Yunfan, CHEN Du, LI Baosheng, ZHENG Yongjun. Measurement Method of Maize Ear Characters Based on Machine Vision[J]. Transactions of the Chinese Society for Agricultural Machinery,2020,51(s2):357-365.

復(fù)制
分享
文章指標(biāo)
  • 點擊次數(shù):
  • 下載次數(shù):
  • HTML閱讀次數(shù):
  • 引用次數(shù):
歷史
  • 收稿日期:2020-08-17
  • 最后修改日期:
  • 錄用日期:
  • 在線發(fā)布日期: 2020-12-10
  • 出版日期: 2020-12-10
文章二維碼