| In recent years,with the continuous development of family breeding in rural pastoral areas,the demand for forage has increased day by day.At the same time,users have put forward clear requirements for the quality of high-quality forage.As a high-quality forage crop,alfalfa is planted in a large proportion in the developed areas of animal husbandry.In order to meet the needs of the industry,research and explore the relationship and application of alfalfa cutting in different growth periods on the microstructure,morphology,quality and comprehensive benefits,Is conducive to improving the accuracy of alfalfa production and management,and is conducive to improving the scientific and technological level of alfalfa industry development.The research uses digital image processing technology to analyze and process its stem anatomical structure images,reveals the anatomical structure and structure of alfalfa stems,and studies and analyzes the correlation between structural indicators and different phenological periods.According to the requirements of plant biology experiment,the stem samples of alfalfa at branching stage,budding stage,blooming stage and blooming stage were collected,and permanent cross sections of the stem were made by paraffin sectioning technique,and the tissue cells were stained by double staining.Combining the biological optical microscope with the computer supporting software Toup View3.7,the anatomical structure images of alfalfa stems were collected.According to the image characteristics of stem anatomy,based on the image color saliency and shape features,pre-segmentation is performed on the cross-section of stem anatomical structure,vascular bundle lignified cells,medullary parenchyma cells and epidermal cells images;use mathematical morphology to process the connection Neighboring objects,smoothing boundaries and removing residual noise;image segmentation algorithms for stem cross-section,vascular bundle lignified cells,marrow parenchyma cells and epidermal cells are evaluated by image pixel indicators.The results show that the mathematical morphology hole filling and connected domain labeling method has a significant effect on the cross-sectional image processing of the stem anatomical structure,and the largest connected area is the target area;the image color feature and shape feature are combined to process the image of vascular bundle lignified cells Compared with the image color saliency segmentation process,its segmentation accuracy is significantly improved;the mathematic morphology hole filling and opening operations are used to process the marrow parenchyma cell image,which effectively removes residual noise,and the target area segmentation is relatively complete;in order to achieve epidermal cells For the accurate segmentation of the adhesion structure,the segmented image will be segmented twice through the iteratively eroded watershed segmentation algorithm,and the segmentation effect is ideal.The evaluation results of image segmentation algorithms for the cross-section of stem anatomical structure,vascular bundle lignified cells,pulp parenchyma cells and epidermal cells show that the average segmentation errors are 2.537%,6.140%,6.197%,and 5.321%,respectively.The research is based on image preprocessing and segmentation processing,and the overall parameterization of the stem anatomical structure image section is carried out;the average measurement accuracy of the proposed stem anatomical structure cross-sectional area(SCA)measurement algorithm reaches 99.383%.The number of vascular bundle lignified cells(NPXVB)was measured using the binary image connected component annotation method,and the measurement accuracy reached 97.278%.A measurement algorithm for the width of primary xylem(WPXVB)and the height of tandem structure(TSH)was proposed by Toupview3.7 The average relative errors of the software manual measurement values were 3.578% and 3.413% respectively.The study separately parameterized the thickness and number of epidermal cells of the stem anatomical structure;through the comparison of interactive measurement and algorithm measurement,the average relative error was 2.689% and 3.034%,respectively.Analysis result: Through research and analysis of the correlation between the anatomical structure of alfalfa stem and the phenological phase,the anatomical structure of the stem in each growth period is observed and compared,and it has strong plasticity;the cross-sectional area of stem anatomy(SCA),dimension The number of vascular bundle lignified cells(NPXVB),vascular bundle primary xylem width(WPXVB),tandem structure height(TSH)and epidermal cell thickness(ECT),etc.have significant changes in different phenological stages,while the number of epidermal cells(NEC)No significant changes were seen in different phenological periods.Studies have shown that the use of digital image processing technology to extract and measure features of alfalfa stem anatomical structure images has the advantages of high operational stability,low labor intensity,high precision and high efficiency;the processing algorithm proposed in this thesis is effective for the anatomy of alfalfa Structural image research provides a new technical means,and lays a good foundation for its quality detection and correlation analysis between environmental factors. |
