Image Analysis And System Realization Of Phenotypic Characteristics Of Millet Root System

Root system is an important organ for crops to obtain water and nutrients needed for growth in soil.It can first respond to various changes in soil and plays an important role in the whole growth process of crops.Millet originated in China,water and fertilizer utilization rate is high,so drought resistance and need less fertilizer and pesticides,can adapt to the drought environment.These characteristics are closely related to the growth and development of root system,so the measurement and analysis of root phenotype is of great significance for millet,a crop with efficient utilization of soil water and nutrients.Traditional manual method to measure crop root phenotype data requires a lot of manpower and material resources and is inefficient,which brings great inconvenience to relevant researchers.With the rapid development of computer vision and digital image processing technology and the improvement of imaging hardware in performance and resolution,it has become possible for computer automatic measurement to replace the traditional manual contact measurement,and the use of machines for detection,identification and measurement analysis has gradually become the mainstream.Compared with the traditional image detection technology,the image detection technology based on computer vision has the advantages of simple operation,instant data acquisition and appropriate accuracy.This study in order to mature after the original dell and two tons of valley 1 millet varieties of root system as the research object,using the mining method to obtain the complete millet roots,washing to remove the root surface sticky soil,according to the section of the root in reverse order to cut out a detailed root every day,to root on the basis of image preprocessing,the thinning algorithm and partial correlation algorithm for section skeleton diagram and the root number,finally by using this method and the manual measurement methods on the roots of the two varieties phenotype measurements and the results are compared.The main results are as follows:1.A method based on pixel scale for automatic screening and root fitting of complex root points,i.e.the intersection of multiple roots,was proposed,with a screening accuracy of 98.0% and a fitting accuracy of 90.0%.2.Based on the traditional thinning algorithm and the local correlation algorithm,the complete root individual and its branches were identified,and the measurement accuracy of the root length reached93.7%.3.The total root length fitting degree and branch number fitting degree of Yuanping cultivars reached93.7% and 99.0%;The fitting degree of the total root length and the number of branches reached 98% and97% respectively.The measured data showed that the system was able to express the difference between the two varieties.4.An automatic measurement system of root phenotype was developed.The functions of the system include image preprocessing,skeleton extraction,branch judgment and phenotypic acquisition.It takes about one minute to measure the root data of a single plant,which overcomes the shortcomings of traditional measurement methods such as time consuming and laborious,low efficiency and large error,and realizes automatic high-throughput and high-precision measurement of millet root phenotype.