Chlorophyll Fluorescence Measurement Of Canopy Based On RGB-D And Cropobserver

Chlorophyll fluorescence measurement is an important means of non-destructive measurement of crop photosynthesis intensity,which can realize real-time and rapid measurement of photosynthetic parameters.At present,the study of chlorophyll fluorescence monitoring on the two-dimensional image level of a single leaf tends to be mature,but the three-dimensional detection of chlorophyll fluorescence of the whole plant and the canopy is still in the exploration stage.In view of the above problems,this paper selects the RGB-D(RGB depth)sensor Kinect and Canopy Chlorophyll fluorescence detection device Cropobserver to study the three-dimensional distribution monitoring of Canopy Chlorophyll fluorescence based on RGB-D and Cropobserver,which provides important technical support for the acquisition and research of 3D comprehensive visualization distribution information of Crop Canopy Chlorophyll fluorescence distribution.The main research contents and results are as follows:(1)Establishment of rgb-d data acquisition experimental platform based on KinectIn order to achieve the simultaneous acquisition of RGB-D image information and Canopy Chlorophyll fluorescence,the color information flow and depth information flow of crop canopy were collected by PC workstation when Crop Canopy Chlorophyll fluorescence information was collected by cropobserver.The matching between depth map and color map is realized by mapping the pixel coordinates of depth map to color map and camera internal parameter model.It provided conditions for the three-dimensional distribution of chlorophyll fluorescence information in plant canopy.(2)Position matching between Kinect and CropobserverIn order to achieve the spatial location matching between Kinect and cropobserver,the RGB-D image of laser measurement point at the edge of the measurement range of cropobserver was captured by Kinect.The spatial location relationship model of Kinect camera and cropobserver was established by using the calibration principle of binocular camera,and the transformation and location matching between the measurement point coordinates of cropobserver and the spatial coordinates in the depth map were realized.It provides the basis for the generation,reconstruction and matching of the three-dimensional morphological point cloud and chlorophyll fluorescence point cloud.(3)Extraction of morphological parameters of plant canopyThe RGB-D image of cucumber canopy was acquired by Kinect overhead photography.The depth map was preprocessed by depth through filtering and condition filtering,and the plants in the depth map were extracted.Through color image and depth image processing,the plant height,crown length and other morphological parameters were extracted and analyzed.Through gray-scale processing,segmentation and denoising of the plant color image,the external rectangle statistics are carried out,combined with the pixel size,the crown length is obtained.The height from Kinect to the center of the plant is obtained by traversing the minimum depth in the depth map of the center of the plant,and the plant height is calculated.The experimental results show that the goodness of fit between cucumber plant height and actual plant height obtained by this method is R~2=0.9998,RMSE=0.219cm,and the goodness of fit between crown length and actual crown length R~2=0.9991,RMSE=0.624cm.The plant height and crown width obtained by this method are suitable for the measurement and analysis of canopy morphology of tall crown plants.(4)Three dimensional characterization of effective chlorophyll fluorescence points obtained by cropobserver in plant canopyAccording to the RGB-D image obtained by Kinect,through color extraction,filtering outliers based on k-nearest-neighbor algorithm,and European clustering segmentation,three-dimensional point cloud of plant canopy is generated.By attaching chlorophyll fluorescence parameters such as photosynthetic efficiency,light intensity,relative electron transport rate and measurement time to the point cloud color space channel,the point cloud distribution characterization and 3D visualization of continuous chlorophyll fluorescence monitoring are realized.The results of sparse measurement and dense measurement of single cucumber and group measurement of cucumber were used to verify the function of the system.The results showed that this method can better realize the three-dimensional distribution rule representation and three-dimensional visualization display of chlorophyll fluorescence parameters of different types of cucumber plants.