农机化研究
農機化研究
농궤화연구
Journal of Agricultural Mechanization Research
2016年
8期
126-130
,共5页
张震%高雄%陈铁英%王海超
張震%高雄%陳鐵英%王海超
장진%고웅%진철영%왕해초
农业自动化%喷雾机械%机器视觉%光谱成像系统%甘蓝夜蛾%自动识别%特征波段
農業自動化%噴霧機械%機器視覺%光譜成像繫統%甘藍夜蛾%自動識彆%特徵波段
농업자동화%분무궤계%궤기시각%광보성상계통%감람야아%자동식별%특정파단
agricultural automation%spraying machine%machine vision%spectral imaging systems%cabbage moth%auto-matic recognition%characteristic wave band
农业喷雾对象的识别和定位是农业自动化喷雾机械研究中的核心技术之一. 对病虫害甘蓝进行精准喷洒农药,实现病虫害准确自动识别成为关键. 为此,利用机器视觉的欧氏距离甘蓝夜蛾虫害自动识别检测系统,结合由Qualityspec 光谱仪组成的光谱成像系统,对甘蓝正常叶片和遭受甘蓝夜蛾虫害的甘蓝叶片的颜色特征和光谱特征进行分析,并采用机器视觉分割阈值选取中的Otsu算法和自适应波段选择方法提取出了颜色差异的最佳几何阈值和两种叶片的特征波段. 试验结果表明:综合机器视觉和光谱技术能够实现甘蓝夜蛾虫害的自动且准确的识别,准确率可达94%. 因此,建立机器视觉和光谱技术综合识别体系,可为农作物病虫害自动防治喷雾机器人的研制奠定基础,以达到农作物病虫害实时识别和及时治理的目的.
農業噴霧對象的識彆和定位是農業自動化噴霧機械研究中的覈心技術之一. 對病蟲害甘藍進行精準噴灑農藥,實現病蟲害準確自動識彆成為關鍵. 為此,利用機器視覺的歐氏距離甘藍夜蛾蟲害自動識彆檢測繫統,結閤由Qualityspec 光譜儀組成的光譜成像繫統,對甘藍正常葉片和遭受甘藍夜蛾蟲害的甘藍葉片的顏色特徵和光譜特徵進行分析,併採用機器視覺分割閾值選取中的Otsu算法和自適應波段選擇方法提取齣瞭顏色差異的最佳幾何閾值和兩種葉片的特徵波段. 試驗結果錶明:綜閤機器視覺和光譜技術能夠實現甘藍夜蛾蟲害的自動且準確的識彆,準確率可達94%. 因此,建立機器視覺和光譜技術綜閤識彆體繫,可為農作物病蟲害自動防治噴霧機器人的研製奠定基礎,以達到農作物病蟲害實時識彆和及時治理的目的.
농업분무대상적식별화정위시농업자동화분무궤계연구중적핵심기술지일. 대병충해감람진행정준분쇄농약,실현병충해준학자동식별성위관건. 위차,이용궤기시각적구씨거리감람야아충해자동식별검측계통,결합유Qualityspec 광보의조성적광보성상계통,대감람정상협편화조수감람야아충해적감람협편적안색특정화광보특정진행분석,병채용궤기시각분할역치선취중적Otsu산법화자괄응파단선택방법제취출료안색차이적최가궤하역치화량충협편적특정파단. 시험결과표명:종합궤기시각화광보기술능구실현감람야아충해적자동차준학적식별,준학솔가체94%. 인차,건립궤기시각화광보기술종합식별체계,가위농작물병충해자동방치분무궤기인적연제전정기출,이체도농작물병충해실시식별화급시치리적목적.
The spray object recognition and localization is one of the core technology of automatic spray mechanization re -search .For precision spraying pesticide plant diseases and insect pests of cabbage , the accurate and automatic identifica-tion of plant diseases and insect pests of cabbage becomes the key .Therefore , using machine vision automatic identifica-tion of the Euclidean distance of cabbage moth pests detection system , combined with spectral imaging system composed of qualityspec spectrometer , Cabbage normal blade and suffer from the cabbage moth pests of cabbage leaf color features and spectral characteristics were analyzed .The best geometric threshold of color difference and characteristic bands of two kinds of leaves were extracted , using the Otsu threshold value image segmentation algorithm and adaptive band selection method.The test results show that the technology compositing image processing with spectrum can realize automatic and accurate identification of Cabbage moth pests , accuracy reaching 94%.Therefore, the establishment of comprehensive image processing and spectroscopy identification system can lay the foundation for the development of the automatic con -trol of spray robot , in order to achieve the purpose of crop diseases and pests identification and real -time treatment .