柑橘叶面病害监测颜色空间改进算法研究
|
摘要
柑橘叶面病害是影响柑橘产量和质量的重要因素,如何对柑橘叶面进行病害识别是后期病害检测的一个关键步骤。为了提高柑橘在各种病害环境下病斑图像识别的准确率,提出了一种针对柑橘病害监测的HSV颜色直方图空间的图像检索改进算法。此算法用颜色空间来展现柑橘病害所导致的颜色变动,根据柑橘叶面病变产生的异常颜色信息,结合传统直方图图像检索方法,对柑橘叶面图像颜色进行非均匀量化处理,且将非均匀化量化处理后的归一化颜色矩阵作为因子来进行图像检索。使用者可以经过使用这种形式来检测柑橘叶面图像,及早发现病害。与传统的方案相比,此算法在柑橘叶面病害监测方面的检索查准率和查全率均有显著提高,从而验证了本文算法的有效性。
Citrus leaf diseases are important factor affecting the yield and quality of citrus. How to find the disease of citrus leaves is a key step in the late disease detection. In order to improve the accuracy of image recognition in orange leaf spots in various disease conditions,a new citrus leaf diseases image retrieval method based on HSV( Hue,Saturation,Value) Histograms is proposed,which by using color space to show the citrus diseases lead color changes,according to the abnormal color of citrus leaf disease information,combined with image retrieval method which based on the traditional histogram,non-uniform quantization of the citrus leaf color image is used in this method,and the heterogeneity of normalized color matrix quantization processed as a factor for citrus leaf image retrieval. The user can use the method to find the Citrus leaf diseases by retrieving images of citrus leaf spots. Compared with the traditional image retrieval methods,the method can not only improve retrieval speed,but also have better results.
Citrus leaf diseases are important factor affecting the yield and quality of citrus. How to find the disease of citrus leaves is a key step in the late disease detection. In order to improve the accuracy of image recognition in orange leaf spots in various disease conditions,a new citrus leaf diseases image retrieval method based on HSV( Hue,Saturation,Value) Histograms is proposed,which by using color space to show the citrus diseases lead color changes,according to the abnormal color of citrus leaf disease information,combined with image retrieval method which based on the traditional histogram,non-uniform quantization of the citrus leaf color image is used in this method,and the heterogeneity of normalized color matrix quantization processed as a factor for citrus leaf image retrieval. The user can use the method to find the Citrus leaf diseases by retrieving images of citrus leaf spots. Compared with the traditional image retrieval methods,the method can not only improve retrieval speed,but also have better results.
引文
[1]王跃勇,于海业,刘媛媛.基于形态学图像检测的机械手移栽穴苗识别技术[J].光学精密工程,2016,24(3):651-658.
[2]汪启伟.图像直方图特征及其应用研究[D].长沙:中国科技大学,2014:1-89.
[3]李丽莎.基于HSV空间的彩色图像亚像素边缘检测[D].西安:西安工业大学,2015:1-71.
[4]张凯兵,章爱群,李春生,等.基于HSV空间颜色直方图的油菜叶片缺素诊断[J].农业工程学报,2016,32(19):179-187.
[5]王长斌.基于云计算的农作物病虫害多源遥感数据挖掘[J].电子技术,2016,3(5):1-17.
[6]濮永仙.基于病斑特征融合的烟草病害图像检索方法[J].河南农业科学,2015,44(2):71-76.
[7]刘梅.基于计算机视觉的玉米种子形态识别测量[J].农机化研究,2018,40(4):210-214.
[8]周彦卿,张卫,顾静良,等.基于HSV空间的红外偏振图像改进融合算法[J].激光与红外,2014,44(12):1379-1383.
[9]于翔,李莹.HSV颜色空间的量化对南通蓝印花布颜色直方图影响的比对研究[J].电子技术与软件工程,2014(15):111-113.
[10]戴佳哲.基于内容的图像检索算法研究[D].杭州:浙江理工大学,2016.
[11]黄冲,刘万才.试论物联网技术在农作物重大病虫害监测预警中的应用前景[J].中国植保导刊,2015,35(10):55-60.
[12]王建文,林劼.基于颜色直方图金字塔的图像自动标注方法[J].计算机工程,2016,42(6):235-240.
[13]颜秉忠.机器视觉技术在玉米苗期杂草识别中的应用[J].农机化研究,2018,40(3):212-216.
[14]谭光兴,刘臻晖.基于改进的排序学习的图片检索算法研究[J].计算机科学,2015,42(12):275-277.
[15]唐翠翠.基于多源遥感数据的小麦病虫害大尺度监测预测研究[D].合肥:安徽大学,2016:1-49.
[2]汪启伟.图像直方图特征及其应用研究[D].长沙:中国科技大学,2014:1-89.
[3]李丽莎.基于HSV空间的彩色图像亚像素边缘检测[D].西安:西安工业大学,2015:1-71.
[4]张凯兵,章爱群,李春生,等.基于HSV空间颜色直方图的油菜叶片缺素诊断[J].农业工程学报,2016,32(19):179-187.
[5]王长斌.基于云计算的农作物病虫害多源遥感数据挖掘[J].电子技术,2016,3(5):1-17.
[6]濮永仙.基于病斑特征融合的烟草病害图像检索方法[J].河南农业科学,2015,44(2):71-76.
[7]刘梅.基于计算机视觉的玉米种子形态识别测量[J].农机化研究,2018,40(4):210-214.
[8]周彦卿,张卫,顾静良,等.基于HSV空间的红外偏振图像改进融合算法[J].激光与红外,2014,44(12):1379-1383.
[9]于翔,李莹.HSV颜色空间的量化对南通蓝印花布颜色直方图影响的比对研究[J].电子技术与软件工程,2014(15):111-113.
[10]戴佳哲.基于内容的图像检索算法研究[D].杭州:浙江理工大学,2016.
[11]黄冲,刘万才.试论物联网技术在农作物重大病虫害监测预警中的应用前景[J].中国植保导刊,2015,35(10):55-60.
[12]王建文,林劼.基于颜色直方图金字塔的图像自动标注方法[J].计算机工程,2016,42(6):235-240.
[13]颜秉忠.机器视觉技术在玉米苗期杂草识别中的应用[J].农机化研究,2018,40(3):212-216.
[14]谭光兴,刘臻晖.基于改进的排序学习的图片检索算法研究[J].计算机科学,2015,42(12):275-277.
[15]唐翠翠.基于多源遥感数据的小麦病虫害大尺度监测预测研究[D].合肥:安徽大学,2016:1-49.