移栽是蔬菜和花卉生產(chǎn)過程中最重要的環(huán)節(jié)之一,為了滿足穴盤苗的生長(zhǎng)空間和養(yǎng)分需求,需將其從高密度穴盤中移栽到低密度穴盤。由于人工幼苗分選的效率較低和勞動(dòng)強(qiáng)度較高,目前難以擴(kuò)大生產(chǎn)規(guī)模。因此,分選出健康幼苗是實(shí)現(xiàn)高質(zhì)量移栽的前提和基礎(chǔ)。另外,植物損傷信息檢測(cè)是穴盤苗自動(dòng)移栽的重要問題。基于圖像處理的檢測(cè)技術(shù)具有速度快、效率高、精準(zhǔn)可靠的優(yōu)點(diǎn),視覺檢測(cè)對(duì)提高效率、減輕勞動(dòng)強(qiáng)度、保證移栽速度、促進(jìn)苗木快速發(fā)展具有重要意義。近年來,已經(jīng)提出了幾種基于圖像的檢測(cè)方法,并且可能將其開發(fā)用于農(nóng)業(yè)生產(chǎn)。然而,現(xiàn)有的基于深度信息的識(shí)別技術(shù)主要是通過深度點(diǎn)云的3D重建,圖像特征提取或RBG信息的融合來實(shí)現(xiàn)幼苗的識(shí)別,復(fù)雜程度高并且對(duì)健康幼苗的識(shí)別非常有限。針對(duì)上述缺點(diǎn),本文提出了利用RealSense深度傳感器開發(fā)了基于深度信息的健康溫室穴盤苗的識(shí)別方法。主要研究?jī)?nèi)容如下:1、根據(jù)移栽要求,研究了黃瓜幼苗形態(tài)特征參數(shù)與苗木之間指標(biāo)的相關(guān)性和灰色關(guān)聯(lián)度分析,確定了黃瓜苗葉面積,莖直徑和株高的閾值。結(jié)論:當(dāng)黃瓜幼苗的葉面積,莖直徑和株高分別大于257mm ²、1mm和27mm時(shí),即可進(jìn)行移...

【文章頁數(shù)】：88 頁

【學(xué)位級(jí)別】：碩士

【文章目錄】：
ACKNOWLEDGEMENT
DEDICATION
ABSTRACT
摘要
ABBREVIATION
CHAPTER 1 INTRODUCTION
    1.1 Global perspective, and Research background
    1.2 Research status of monitoring of seedling in automatic transplanting
        1.2.1 Research status of seedlings transplanter based on MV in abroad
        1.2.2 Research status of seedling transplanter based on MV in P.R. China
    1.3 Research status of MV technology for seedlings recognition at home andabroad
        1.3.1 Research status of MV technology for seedlings recognition used inautomatic transplanting
        1.3.2 Research status of MV technology for seedlings recognition based onimages
    1.4 Research status of MV technology for recognition of fruits, vegetables andplants based on depth vision
        1.4.1 Time-of-flight (ToF) Camera
        1.4.2 Laser Range Finder
        1.4.3 RGB-D Camera
    1.5 Problem statements
    1.6 Research objectives
    1.7 Outline of the dissertation
CHAPTER 2 Correlation analysis of morphological characteristics parameters of cucumberseedlings based on the seedling index
    2.1 Introduction
    2.2 Correlation analysis between the seedling index and morphologicalcharacteristics parameters of cucumber Seedlings
        2.2.1 Experimental setup and seedlings growth
        2.2.2 Measurement of plant morphological characteristics parameters
        2.2.3 Statistical analysis
        2.2.4 Results
            2.2.4.1 Correlation analysis between the seedling index and morphological characteristics parameters of cucumberseedlings
            2.2.4.2 Grey correlation analysis between the seedling index and morphological characteristics parameters of cucumberseedlings
    2.3 Determination of image recognition characteristics of Cucumber seedling
        2.3.1 Determination of image recognition characteristics of Cucumberseedling
        2.3.2 Determination of threshold values for the image recognition ofcucumber seedlings
CHAPTER 3 Seedling-lump integrated monitoring and morphological characteristicsextraction with Real Sense sensor
    3.1 Introduction
    3.2 Materials and Methods
        3.2.1 Materials
            3.2.1.1 Selection of image acquisition equipment
                3.2.1.1.1 Realsense technology introduction
                3.2.1.1.2 Realsense 3D camera parameter comparison
        3.2.2 Methods
            3.2.2.1 Monitoring strategy and overview of the proposedalgorithm
            3.2.2.2 Calibration of the Real Sense sensor
        3.2.3 Close-shot monitoring of seedling with Real Sense sensors
        3.2.4 Image acquisition with intel Real Sense Sensor
            3.2.4.1 RGB image acquisition
            3.2.4.2 Depth image acquisition
        3.2.5 Elimination of background and segmentation
        3.2.6 Plant morphological features extraction with an MV system
            3.2.6.1 Stem diameter
            3.2.6.2 Plant height
            3.2.6.3 Leaf area
            3.2.6.4 Perimeter
        3.2.7 Monitoring of seedling in multi-views
        3.2.8 Monitoring of seedling in different light conditions
    3.3 Experiment to validate the Real Sense sensor
    3.4 Data analysis
    3.5 Results and discussions
        3.5.1 Monitoring morphological characteristics of the seedlings
        3.5.2 Multi-views performances of seedling height
        3.5.3 Seedlings monitoring in different light conditions
        3.5.4 Validation of the sensor
        3.5.5 Impact of multi-views-based comparison
        3.5.6 Influence of light condition
        3.5.7 Reliability of Real Sense sensor for monitoring
        3.5.8 Close-shot monitoring of seedling
CHAPTER 4 Test and analysis of healthy seedlings recognition using functional automatictransplanting Machine system
    4.1 Introduction
    4.2 Brief introduction of full-automatic integrated transplantingmachine system
        4.2.1 Structure composition of transplanting machine system
        4.2.2 Working principle of transplanting machine system
    4.3 Design of image recognition scheme for healthy seedlings based onReal Sense camera
        4.3.1 Seedlings recognition strategy on transplanting machine system
    4.4 Feasibility verification test of healthy seedlings recognition scheme
        4.4.1 Test materials and methods
            4.4.1.1 Test materials
            4.4.1.2 Test methods
        4.4.2 Results and analysis
CHAPTER 5 Conclusion and future prospects
    5.1 Conclusion
    5.2 Existing problems and future prospects
Publication



本文編號(hào)：3992195