一种基于聚类特征的Faster R-CNN粮仓害虫检测方法

本文基于Faster R-CNN模型提出复杂背景下粮仓害虫的检测识别方法。将六种常见的储粮害虫(豆象、谷蠹、米象、锯谷盗、赤拟谷盗、锈赤扁谷盗)分别以大米、小米为背景,建立了真实背景下粮仓害虫图像数据集SGI-6。SGI-6中包括网络获取图像、显微镜采集图像和单反拍摄图像三种多目标尺度的数据集。根据粮仓害虫的小目标特性,使用聚类算法改进Faster R-CNN模型的区域提案网络,来提取这些图像中含有害虫的区域,并对这些区域中的害虫进行分类。实验结果表明,该方法能够在储粮条件下检测和识别粮仓害虫,且其平均准确率(mAP)达到96.63%。

A Faster R-CNN Method for Insect Detection in Stored Grain Based on Clustering Feature

Based on the Faster R-CNN model, this paper proposes a method for detecting and identifying granary insects under complex background. Six common stored grain insects (bean weevils, rhizopertha, sitophilus oryzae, oryzaephilus surinamensis, tribolium castaneum, cryptolestes ferrugineus) were established with rice and millet as the background, and the image data of granary pests under real background was established. Set SGI-6. The SGI-6 includes three multi-target scale datasets for network acquisition images, microscope acquisition images, and SLR images. According to the small target characteristics of granary insects,the clustering algorithm is used to improve the regional proposal network of the Faster R-CNN model to extract the pest-containing areas of these images and classify the insects in these areas. The experimental results show that the method can detect and identify granary insects under the condition of grain storage, and its average accuracy (mAP) reaches 95.63%.