基于自适应滤波算法的树木年轮测量方法  

Annual-ring Measurement Method Based on Adaptive Filtering Algorithm

作　　者：姚建峰 赵燕东[3] 卢军[1] 郑一力[3] 高瑞东[4] 唐守正 YAO Jianfeng;ZHAO Yandong;LU Jun;ZHENG Yili;GAO Ruidong;TANG Shouzheng(Research Institute of Forest Resources Information Techniques, Chinese Academy of Forestry, Beijing 100091, China;College of Computer and Information Technology, Xinyang Normal University, Xinyang 464000, China;School of Technology, Beijing Forestry University, Beijing 100083, China;Guancenshan State-owned Forest Administration of Shanxi Province, Xinzhou 034000, China)

机构地区：[1]中国林业科学研究院资源信息研究所,北京100091 [2]信阳师范学院计算机与信息技术学院,信阳464000 [3]北京林业大学工学院,北京100083 [4]山西省管涔山国有林管理局,忻州034000

出　　处：《农业机械学报》2020年第3期216-222,共7页Transactions of the Chinese Society for Agricultural Machinery

基　　金：中央级公益性科研院所基本科研业务费专项资金重点项目(CAFYBB2018SZ007);国家重点研发计划项目(2017YFD0600403)。

摘　　要：为提高树木年轮测量仪的测量精度,提出了一种自适应低通滤波算法对年轮测量仪的直流电机电流信号进行滤波处理。普通低通滤波算法滤波系数固定,当树木年轮宽度发生变化时,既可能保留了部分噪声信号,又可能滤去部分有效的树木年轮信号,因此降低了年轮测量仪的测量精度。自适应低通滤波算法的滤波系数可随滤波器输入值和输出值的变化自动进行调整,提高了噪声滤除效果,减少了有效信号损失。使用自主研发的年轮测量仪测量了大、中、小3种径阶的60个落叶松圆盘,记录直流电机原始电流信号,分别采用自适应低通滤波算法和普通低通滤波算法对原始电流进行滤波处理,根据滤波后的电流波形图,预估圆盘年龄,并计算2种滤波算法的年轮识别正确率。结果表明,自适应低通滤波器年轮识别正确率的平均值为89.66%,普通低通滤波器年轮识别正确率的平均值为75.84%;当平均年轮宽度变窄时,自适应低通滤波器的年轮识别正确率从95.61%下降到82.02%,下降了13.59个百分点,而普通低通滤波器的年轮识别正确率从86.37%下降到62.30%,下降了24.07个百分点。表明自适应低通滤波器年轮识别正确率高,自适应能力强,年轮识别精度稳定。In order to improve the accuracy of annual-ring measuring instrument,a new adaptive low-pass filtering algorithm was proposed to filter the direct current(DC)motor current signal of annual-ring measuring instrument.The filter coefficients of ordinary low-pass filtering algorithm were fixed.When the annual-ring width was changed,the current fileted by ordinary low-pass filter may not only retain part of the noise signal,but also filter part of the effective annual-ring signal,which reduced the accuracy of the annual-ring measuring instrument.The filter coefficients of the adaptive low-pass filter algorithm can be adjusted automatically with the changes of the input and output values of the filter,which can improve the effect of noise filtering and reduce the loss of effective signals.Totally 60 larch discs,including large,medium and small diameter orders were measured by self-manufactured annual-ring measuring instrument.The original current signals of DC motor were recorded.The adaptive low-pass filtering algorithm and the ordinary low-pass filtering algorithm were used to filter the original current signals of DC motor respectively.According to the filtered current waveform,the disc age was estimated,and the annual-ring recognition accuracy of the two filtering algorithms was calculated.The average annual-ring recognition accuracy of adaptive low-pass filter was 89.66%,and that of ordinary low-pass filter was 75.84%.The average annual-ring recognition accuracy of adaptive low-pass filter was 13.82 percentage points higher than that of ordinary low-pass filter.When the average ring width became more narrow,the annual-ring recognition accuracy of the adaptive low-pass filter was decreased by 13.59 percentage points from 95.61%to 82.02%,while that of the ordinary low-pass filter was decreased by 24.07 percentage points from 86.37%to 62.30%.The experiment results showed that the adaptive low-pass filter had high annual-ring recognition accuracy,strong self-adaptive ability and stable annual-ring recognition accuracy

关 键 词：年轮测量仪 电流滤波 低通滤波器 自适应滤波器 

分 类 号：S758.7[农业科学—森林经理学]