Non-destructive Detection Of Plant Roots Based On Magnetic Resonance Imaging Technology

The information acquisition of plant growth and diagnosis of the growth condi-tion is an important part of the implementation of precision agriculture technology. Plant roots system is a vital organ to adapt to the living conditions, the ability of plants to absorb nutrients is closely related to the root morphology configuration space conditions in the soil. Different root architecture significantly affect the nutrient absorption and utilization of water, carbon allocation and the ability to adapt to envi-ronmental stress and other important physiological functions. In addition, the distribu-tion of soil nutrients or moisture also play some roles in regulating control action on plant root morphology and architecture. However, due to the opacity of the root growth environment, the limitation of their own complex morphology and the quanti-tative analysis methods, the study of root morphology configuration is lagging far from being able to meet the research needs of modern biological roots. In this thesis, using corn as the main research object, we developed the nuclear magnetic resonance imaging, digital image processing technology and computer graphics technology, recognition of root morphology was established, including the establishment of opti-mal conditions for root NMR experiment, interactive3D model and measuring the root form of quantitative analysis of the configuration parameters, the gradient of ni-trogen, phosphorus, quantitative root morphology configuration stress analysis were completed, the maize root morphology and architecture in situ non-destructive testing was realized. The main contents and innovative achievements are as follows:(1) The optimal experimental conditions roots of NMR were established, and a few basic questions about the crop root NMR were studied including three aspects such as the growth medium, magnetic resonance imaging and parameters of root sys-tem which affect the analysis of the quality of magnetic resonance imaging roots fac-tors. The quality and cross-over factors effecting MRI analysis of the root system were analyzed and summarized,21test samples was prepared to conduct the root magnetic resonance imaging test. The optimal MRI condition of maize roots is:3.0T main magnetic field strength+maize root system+ordinary soil+35%soil volumet-ric water content.(2) A silhouette image de-noising algorithm based wavelet transform was es-tablished. MRI images for weak root system noise, wavelet de-noising algorithm and contour wave de-noising algorithm has better de-noising effect, the PSNR is greater than88, preferably retained the of valid information of images.For strong roots MRI image noise, denoising effect wavelet de-noising algorithm and contour wave de-noising algorithms were all decreased, the PSNR dropped to about82. Noising algorithm and the minutiae image retention are superior to the ef-fect of contour wave wavelet algorithms de-noising effect, although the overall smoothness of the wavelet de-noising algorithm is better, but minutiae image severely damaged and the detail of the roots edge are blurred, the targets have smaller rhizome root value were weakened or eliminated, partial distortion.(3)3D segmentation algorithm based on the geometric characteristics of crop roots and their spatial configuration knowledge has been proposed. Among the use of magnetic resonance image of sequences roots, particularly between its upper and lower fault data attributes of the image (such as gray value) and root branching spatial properties (such as location coordinates) showed a regular change of this characteris-tic, established roots segmentation constraints. From maize roots rhizome level, adopting the recursive partitioning drill ideas, based on the fuzzy segmentation roots, a comprehensive image segmentation method based on the geometric characteristics of the root system was designed, and false imagines of the pseudo root zone axes in image layers were cleared to realize the precise three-dimensional magnetic resonance image sequence segmentation of roots.65consecutive MRI sequence images of root system for test results showed the good segmentation results obtained three-dimensional image segmentation method based on the geometric characteristics of root comprehensive information. MR slice on the root of the root of the total loss rate Lop and the size of the pixel error rate EoA is small, except slice16#,22#and29 #, average error rate of root pixel loss rate Lop and size of the pixel error rate EoA are4.1%and1.6%, respectively.(4) Proposed a new evaluation system root configuration parameters to achieve a three-dimensional geometry of plant roots type of comprehensive quantitative de-scription. For three nitrogen gradient, three phosphorus stress gradient of corn plants research using visualization techniques to achieve the in-situ roots of non-destructive testing, using the constructed root configuration parameter evaluation system for quantitative analysis. The results showed that the maize root system with high nitro-gen treatment has the shallow root distribution of architecture, the low total root sys-tem, the spatial well-distribution characteristics of the branch root. For the average horizontal root, root difference between each sample under nitrogen stress gradient has little differences, its mean value of the horizontal root is between42.29°-47.68°. For phosphorus stress treatment, high phosphate processing corn root system config-uration in roots exhibited a shallow root distribution, multi-branched roots, total root system characteristics. Similar to root morphology, phosphorus, low nitrogen stress the results showed a lot of configuration for the root elongation product of photosyn-thesis and expansion. In multi-phosphorus stress gradient, the difference between the average horizontal maize root significantly. In the high phosphorus stress the mean horizontal value of the maize root is about48.62°. The circumstances under low phosphorus stress mean value of the root horizontal is about59.72The research results achieved in situ maize root morphology of non-destructive testing and quantitative analysis of configuration parameters, and the further devel-opment of rational fertilizer irrigation strategies provided the theoretical support for improving water use efficiency of crop nutrients which has an important significance.