Research On Nondestructive Detection Of Facility Crops Growth Information And Development Of The Equipment

Facilities agriculture is gradually becoming a hot area in China.Due to the current low level of facilities agriculture,the demand for improving agricultural production efficiency and controlling agricultural production pollution is increasing.Non-destructive testing technology is fast and accurate,can reduce production costs,improve production efficiency and avoid the time-consuming and laborious chemical methods and the disadvantages of experts’manual detection of subjective initiative leading to large errors.It has become a key and effective method for crop detection.The crops grown in this study were tomato and bell pepper,and the non-destructive testing of the facility crops under nutrient and water stress conditions was performed through hyperspectral imagery and machine vision techniques.In addition,based on the limited number of devices currently used in the field of non-destructive testing,a new non-destructive testing device for facility crops was designed using CAE finite element analysis technology.This study mainly completed the following work:（1）Experiments on nondestructive testing methods were conducted.Tomato samples and sweet pepper samples were cultivated by soilless culture and subjected to nutrient stress.Chemical methods were used to measure the true values of crop nitrogen,phosphorus and water content.When used,hyperspectral imagers were used to collect the data of crop leaves under different nitrogen phosphorus conditions in the near-infrared region,including the hyperspectral near-infrared images and the average reflectance of the spectra.The pipeline growth information detection system was used to take pictures of the growing crops at different growth stages.（2）Processing and analysis of hyperspectral information and visual image information.The obtained spectral reflection data is processed to establish a primary regression band of the linear regression equation,and an uncorrelated method is used to perform collinearity processing on the spectral data to remove the collinearity,and finally determine the three characteristic wavelengths of the nitrogen near-infrared waveband as B₁₄₃（1376nm）,B₂₃₆（1688nm）,B₂₅₂（1750nm）,the three characteristic wavelengths of phosphorus near infrared band are B₂₀₄（1573nm）,B₂₁₂（1601nm）,B₂₂₉（1662nm）,and the characteristic wavelength of moisture near infrared band is B₄₈（1061nm）,B₆₀（1104nm）,B₆₇（1129nm）,and a prediction model of nitrogen phosphorus and moisture was established.The noise reduction,color space conversion,binary binarization,and extraction calculations are performed on the crop top view and the main view image,and the data of the crown width,plant height,and stem growth characteristics of the crop are obtained,and at the same time the prediction of the growth potential characteristics is established.model.（3）Comprehensive characteristics modeling of growth potential characteristics and nutritional characteristics.Analyze the growth of tomatoes and sweet peppers.The integrated model of the long-term features and the reflection intensity at the characteristic wavelengths was used as an independent variable.The results showed that the overall effect of the integrated model was better,and the prediction model built with respect to the single long-term feature or single nutrient feature wavelength had better prediction effect,indicating that the comprehensive modeling effect was significant.（4）Suspended-rail type greenhouse environment and crop growth information detection system was designed.The device includes a track beam assembly,a traveling mechanism,a sliding platform,a multi-sensor system,and a control cabinet assembly.Several kinds of structural design schemes are proposed in advance and the final scheme is determined in conjunction with CAE finite element technology.The accuracy of the simulation and the performance of the device were verified through experiments,indicating that the device can accurately capture the environment,crop growth and nutrition information in a greenhouse environment.