摘要
荒漠灌木生物量估测是研究荒漠生态系统结构与功能的重要指标。为了实现荒漠灌木地上生物量的野外实时、无损、快速估测,应用机械设计与制造技术和方法,研制了一种荒漠灌木地上生物量的野外测量装置,综合应用数字图像处理算法,基于Android平台开发了测量装置上运行的荒漠灌木地上生物量无损快速估测软件。以新疆古尔班通古特沙漠南缘的150株柽柳为试验对象,验证荒漠灌木地上生物量无损快速估测方法的有效性。结果表明,以柽柳图像俯视面积(S)为自变量建立生物量模型B=15.215+369.44S-42.03S^2+1.668S^3最优,决定系数(R^2)最大,与人工实际测量结果相比,该系统估测的平均绝对误差为79.12g,平均系统误差为1.73%,平均绝对百分误差为18.59%,估测值和实测值相关系数为0.939 1,能有效测量生物量,为无损快速估测荒漠柽柳地上生物量提高了工作效率。
Estimation of desert shrub biomass is an important index to study the structure and function of desert ecosystem. In order to realize the real-time,nondestructive and rapid estimation of aboveground biomass of desert shrubs,by using the technology of mechanical design and manufacture,a field measuring device for the aboveground biomass of desert shrubs was developed. With the integrated application of digital image processing algorithm based on the Android platform,the non-destructive and rapid estimation software for the aboveground biomass of desert shrubs was developed. The southern fringe of Gurbantunggut Desert in Xinjiang in 150 strains of Tamarix ramosissima as the experimental object,the biomass was effective nondestructive method for fast estimating verification on desert shrubs. The results showed that the optimal biomass model was B = 15. 215 + 369. 44 S-42. 03 S^2+ 1. 668 S^3,which was based on the area of vertical view of Tamarix ramosissima. The coefficient of determination was the maximum val-ue. Compared with the actual measurement results,the average absolute error of the system was 79. 12 g,the average system error was 1. 73%,the mean absolute percentage error was about 18. 59%,and the correlation coefficient between the estimated value and the measured value was about 0. 939 1. It can effectively measure the biomass,improving work efficiency for rapid and non-destructive estimation of aboveground biomass of Tamarix desert.
出处
《浙江农业学报》
CSCD
北大核心
2017年第11期1920-1929,共10页
Acta Agriculturae Zhejiangensis
基金
新疆维吾尔自治区研究生创新项目(XJGRI2015087)
2014年库尔勒科技合作项目(KRLKJHZXM)
关键词
ANDROID平台
柽柳
生物量
无损估测
Android platform
Tamarix ramosissima
biomass
non-destructive estimation