多信息融合的马尾松苗木静态指标检测方法  

Study on static index detection method of masson pine seedlings based on multi-information fusion

作　　者：李玉荣[1] 刘英[1] 汪力 LI Yurong;LIU Ying;WANG Li(College of Electronic and Mechanical Engineering,Nanjing Forestry University,Nanjing 210037,China)

机构地区：[1]南京林业大学机械电子工程学院,南京210037

出　　处：《林业工程学报》2019年第5期129-133,共5页Journal of Forestry Engineering

基　　金：国家自然科学基金(31570714)

摘　　要：获取马尾松形态指标和主要生理指标有助于筛选优良的造林苗木。利用先进的立体视觉3D重构技术和化学计量学方法(近红外光谱检测技术)实现马尾松苗木的静态指标精准测量。首先搭建软硬件实验平台,通过图像采集系统获取图像序列,在计算机上通过算法重构出马尾松苗木的3D模型,提取马尾松苗木的形态学指标;然后利用化学计量学方法挖掘近红外光谱中的有用信息,并采用偏最小二乘回归方法建立线性模型,从而达到苗木主要生理指标的快速、无损测定。所选取的研究方法得到的马尾松苗木质量指标与人工实际测量得到的数据具有很强的相关性。基于多信息融合的马尾松苗木质量检测方法精度高,可满足林业调查要求,并为马尾松苗木活力相关研究提供了新的思路。As one of the main forestation species,masson pine is widely distributed in southern China.Obtaining masson pine morphological index and physiological index is helpful for screening good forestation seedlings.Morphological indicators reflect the external characteristics of masson pine,and physiological indicators reflect the internal physiological status of masson pine.The traditional method of masson pine testing is complicated,time?consuming and laborious.It is of great significance for the construction of modern forestry ecological civilization to quickly and accurately evaluate the quality of masson pine seedlings by means of advanced technologies.A new method based on advanced stereo vision 3D reconstruction technology and the stoichiometric method(near infrared spectroscopy)detection technology to evaluate the quality of masson pine was proposed in this paper.Firstly,the software and hardware platform were built to reconstruct the three?dimensional model of masson pine seedlings by multi view image sequence,and the morphological parameters of the seedlings were extracted quickly.Then the useful information in the near infrared spectrum was excavated by the chemometrics method and the spectral model was established to achieve the rapid and nondestructive measurement of the physiological indexes of the seedlings.The algorithm could optimize the characteristic wave number of near infrared spectrum and improve the accuracy of measurement.The software extracted values and manually measured values of crown,trunk and root length and volume of masson pine seedlings were linear fitted.The quality index of masson pine seedlings selected by the research method was highly correlated with the data obtained by manual measurement.The results showed that thequality detection method of masson pine seedlings based on multi information fusion could meet the requirements of forestry investigation,save time and improve the accuracy,and provide a new way of thinking for the study of the vitality of masson pine.Based on this stud

关 键 词：马尾松 形态学指标 苗木含水量 多信息融合 三维重建 近红外光谱 

分 类 号：S126[农业科学—农业基础科学] S758