中国农业科学院农业信息研究所 中国农学会
Agricultural Information Institute of CAAS China Association of Agricultural Science Societies
En
Ch
中国农业科学院农业信息研究所 中国农学会 主办

AI检索

订阅

返回顶部
当前位置: 首页 > 文章 > 基于改进PatchSVDD的田间异常区域检测方法 河北农业大学学报 2024,47 (1) 106-114
Position: Home > Articles > Farmland anomaly detection based on improved PatchSVDD Journal of Agricultural University of Hebei 2024,47 (1) 106-114

基于改进PatchSVDD的田间异常区域检测方法

作  者：
陈祖强;庞立欣;郭娜炜;蔡金金;么炜;刘博
单  位：
河北农业大学机电工程学院;河北农业大学科学技术研究院;河北农业大学信息科学与技术学院
关键词：
农田监测;异常检测;无人机遥感;三元损失函数;核心集
摘  要：
利用无人机遥感技术对农田进行监测并及时发现田间异常对保证农业生产具有重要意义.目前田间异常区域检测需要标注大量的正常与异常样本.然而,异常样本在整个农田区域中占比过小且无法充分收集.特别是农田异常的多样性和不可预知性,对检测方法的适用性提出了更高的要求.针对以上问题,本文提出基于改进PatchSVDD(Patch-level Support Vector Data Description)模型的田间异常区域检测方法,该方法仅使用田间正常区域的标注信息,即可对田间异常区域进行检测和定位.首先,改进方法引入不相邻图像块之间的边界损失函数,从而提升了正常与异常样本边界的判别性,改进了检测的鲁棒性;第二,引入外部记忆组件,通过压缩存储正常样本特征,从而在保证检测精度的基础上有效减少了测试阶段的时间和空间消耗;第三,构建了包含杂草簇、种植缺失、障碍物、双倍种植和积水共 5 个异常类型的田间异常数据集.本文方法在平均检测AUC(Area Under Curve)值和平均定位AUC值上分别达到了 96.9%和 94.6%,相比于原算法分别提升 1.2%和1.6%,从而验证了方法的有效性.
译  名：
Farmland anomaly detection based on improved PatchSVDD
关键词：
field monitor%anomaly detection%UAV remote sensing%triplet loss function%the core set
摘  要：
It is of significant importance that UAV remote sensing technology was adopted to identify farmland anomalies to ensure agricultural production.Most current farmland anomaly detection methods require labeling a large number of normal and abnormal samples.However,the abnormal samples in the farm area are too small to be collected adequately.In particular,the diversity and unpredictability of farmland anomalies require advanced detection methods.To address these issues,this paper proposed an improved PatchSVDD(Patch-level Support Vector Data Description)model for detecting abnormal areas in the farm,which only utilized the labeling information of normal areas in the farm without labeling the abnormal ones.First,the improved method introduced a margin-based loss function between non-adjacent image patches to improve the discriminability of the boundary between normal and abnormal samples and enhance the robustness of the detector.Second,an external memory module was adopted to store the compressed regular sample features to effectively reduce the time and space consumption in the testing phase while ensuring detection accuracy.Third,a farmland anomaly detection dataset was constructed containing a total of five anomaly types,i.e.,weed clusters,missing planting,obstacles,double planting and standing water.The proposed method achieved 96.9%and 94.6%in the average detection AUC(Area Under Curve)value and average localization AUC value,respectively,which demonstrated improvements of about 1.2%and 1.6%compared with the original PatchSVDD,suggesting the effectiveness of the method.

相似文章

导出引用
计量
文章访问数: 7
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊

友情链接Friendly link
农业信息研究所Agricultural information institute 农业智能知识服务平台Agricultural intelligent knowledge service platform 农业学术服务平台Agricultural Knowledge Service System 中国知网CNKI 维普网VIP 万方WANFANG DATA
联系我们Contact us

电话Tell：010-82109652-808

邮箱E-mail：agrijournal@caas.cn

服务Service
关于我们About us 意见建议Opinions and suggest
版权说明Copyright
单位：中国农业科学院农业信息研究所 中国农学会 © 2017中国农业科学院农业信息研究所 ICP备案号：京ICP备09089781号-46