复杂背景下茶树害虫图像计数算法研究

doi:10.14088/j.cnki.issn0439-8114.2021.20.032

湖北农业科学 ›› 2021, Vol. 60 ›› Issue (20): 158-163.doi: 10.14088/j.cnki.issn0439-8114.2021.20.032

复杂背景下茶树害虫图像计数算法研究

黄灿灿, 陈亚龙, 张伟, 蔡强

浙江清华长三角研究院,浙江嘉兴 314006

收稿日期:2021-06-16 出版日期:2021-10-25 发布日期:2021-11-05
通讯作者: 张伟（1980-）,男,山东沂南人,讲师,博士,主要从事物联网与智能硬件和农业智能装备研究。
作者简介:黄灿灿（1986-）,女,浙江龙游人,副研究员,硕士,主要从事现代化分析测试技术及科学仪器制造研究,（电话）18857372322（电子信箱）huangcancan0105@163.com。
基金资助:
浙江省公益技术研究计划（LGN19F020001）

Research on the algorithm counting tea pests by image in complex background

HUANG Can-can, CHEN Ya-long, ZHANG Wei, CAI Qiang

Yangtze Delta Region Institute of Tsinghua University,Jiaxing 314006,Zhejiang,China

Received:2021-06-16 Online:2021-10-25 Published:2021-11-05

摘要/Abstract

摘要： 依托南方地区茶树田间常用的某型号捕虫设备,采集害虫样本图像,研究基于图像处理技术的害虫计数方法。为了提取样本图像中害虫区域图像,通过几种常见的分离背景方法的试验,如颜色直方图分割和灰度阈值分割,得出样本图像中害虫区域难以分离的原因,进而提出一种网格划分图像的处理方法,即将图像按10×10划分,对划分后的图像做处理以提取害虫区域,该方法能够将害虫区域从背景中完整分离出来;针对图像中害虫区域存在粘连影响害虫计数准确率的问题,提出基于标记控制的分水岭分割算法,利用扩展极小值的方法建立标记,进而完成对粘连区域分割与害虫计数,并进行了噪声测试与害虫计数试验。结果表明,针对该型号捕虫设备采集的样本图像,害虫计数的平均准确率为91.8%,该算法能够完整地提取害虫区域,有效分离粘连重叠的害虫,适用于复杂背景下茶树害虫图像计数。

关键词: 茶树害虫, 分离背景, 粘连, 标记控制, 计数

Abstract: Based on a kind of insect trap lamp equipment, the sample images gained and the counting algorithms with image processing were addressed. Adopted some common algorithms of separating background areas, such as color histogram segmentation and gray threshold segmentation, the sample images were difficult to be separated. This paper develops an image processing algorithm which depend on grid division. Segmented the image as 10×10, and then extracted the pests areas from the separated images. For the overlapping areas will affect the accuracy of counting, developed the marker-controlled watershed segmentation algorithm. Marking the image with minimum extension transform to segment the overlapping areas and then realizing the pests counting, the noise test and counting test were developed furtherly. Experimental results showed that, for the sample images which collected from this kind of pests trap equipment, the average accuracy of the algorithm was 91.8%. This algorithm could extract the pests areas from the sample image completely and segment the overlapped pests effectively. And the algorithm was applicable for tea pests counting in complex background.

Key words: tea pests, image processing, overlap, background area separation, counting

中图分类号:

黄灿灿, 陈亚龙, 张伟, 蔡强. 复杂背景下茶树害虫图像计数算法研究[J]. 湖北农业科学, 2021, 60(20): 158-163.

HUANG Can-can, CHEN Ya-long, ZHANG Wei, CAI Qiang. Research on the algorithm counting tea pests by image in complex background[J]. HUBEI AGRICULTURAL SCIENCES, 2021, 60(20): 158-163.

参考文献

[1] 金星. 光纤发光诱虫灯对茶树害虫的诱捕效果研究[J]. 农机化研究, 2016, 38(7):224-228.
[2] 于新文, 沈佐锐. 棉铃虫数字图像增强技术的研究[A]. 张广学. 走向21世纪的中国昆虫学——中国昆虫学会2000年学术年会论文集[C].北京:中国科学技术出版社, 2000,468-474.
[3] 姚青, 吕军, 杨保军, 等. 基于图像的昆虫自动识别与计数研究进展[J]. 中国农业科学, 2011, 44(14):2886-2899.
[4] EBRAHIMI M A,KHOSHTAGHAZA M H,MINAEI S,et al.Vision-based pest detection based on SVM classification method[J]. Computers and electronics in agriculture, 2017, 137:52-58.
[5] DING W, TAYLOR G.Automatic moth detection from trap images for pest management[J]. Computers and electronics in agriculture, 2016, 123:17-28.
[6] 肖德琴, 张玉康, 范梅红, 等. 基于视觉感知的蔬菜害虫诱捕计数算法[J]. 农业机械学报,2018, 49(3):51-58.
[7] 陈树越, 吴正林, 朱军, 等. 基于凹点检测的粮仓粘连害虫图像分割算法[J]. 计算机工程,2018,488(6):219-224.
[8] XIA C, CHON T S, REN Z, et al.Automatic identification and counting of small size pests in greenhouse conditions with low computational cost[J]. Ecological informatics, 2015, 29:139-146.
[9] 李文勇, 李明, 钱建平, 等. 基于形状因子和分割点定位的粘连害虫图像分割方法[J]. 农业工程学报,2015,31(5):175-180.
[10] WANG Z B, WANG K Y, et al.A cognitive vision method for insect pest image segmentation[J]. International federation of automatic control, 2018, 51(17):85-89.
[11] 杨国国,鲍一丹,刘子毅. 基于图像显著性分析与卷积神经网络的茶园害虫定位与识别[J]. 农业工程学报,2017(6):163-169.
[12] 李文勇. 基于机器视觉的果园性诱害虫在线识别与计数方法研究[D]. 北京:中国农业大学, 2015.
[13] VINCENT L,SOILLE P.Watersheds in digital spaces: An efficient algorithm based on immersion simulations[J]. IEEE transactions on pattern analysis and machine intelligence, 1991, 13(6):583-598.
[14] 王燕. 特定类的图像语义分割[D]. 上海:上海师范大学, 2012.
[15] SOILLE P.Morphological image analysis: Principles and applications[J]. Sensor review, 1999, 28(5):800-801.
[16] 缪慧司,梁光明,刘任任,等. 结合距离变换与边缘梯度的分水岭血细胞分割[J]. 中国图象图形学报,2016,21(2):192-198.
[17] 余旺盛,侯志强,王朝英,等. 基于改进滤波和标记提取的分水岭算法[J]. 电子学报,2011,39(4):825-830.

复杂背景下茶树害虫图像计数算法研究

Research on the algorithm counting tea pests by image in complex background

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 4

编辑推荐

Metrics

本文评价

[1]	康世英, 姚斌. 基于欧拉数的谷物颗粒图像目标计数方法[J]. 湖北农业科学, 2023, 62(8): 197-201.
[2]	庞云娟, 庞兰英, 曾金, 刘康连, 梁晓玲, 龙文洲, 肖萍. 岩黄连栓微生物限度检查方法的适用性试验[J]. 湖北农业科学, 2023, 62(1): 162-166.
[3]	刘保平. 日用洗涤剂对土壤微生物浓度的影响[J]. 湖北农业科学, 2022, 61(16): 40-46.
[4]	郝王丽, 尉培岩, 韩猛, 张丽, 席瑞泽. 基于YOLOv3网络的小麦麦穗检测及计数[J]. 湖北农业科学, 2021, 60(2): 158-160.