文档介绍：: .
中国粮油学报进掩膜预测得更加精细。实验结果表明，该网
络可以实现对粘连米粒的实例分割，与 Mask R-CNN 网络相比，改进后的网络模型
RiceInstNet 的平均精度和召回率分别由 %、%提升到 %、%，同时本网络模
型更轻量，非常适合集成到移动终端或嵌入式设备中。
关键词：实例分割；轻量级；粘连；边界子网络；深度学****br/> 中图分类号：TP29 文献标识码：A
Research on Instance Segmentation Algorithm of Adhesive Rice Grains Based on Deep Learning
Shang Yuting，Wang Yue，Liu Bin
(College of Information and Electronic Engineering , Zhejiang Gongshang University , Hangzhou
310018)
Abstract: Crop appearance quality detection based on machine vision has attracted more and more
attention in recent years. On the sampling inspection table, because rice grains may touch and
adhere to each other, if the collected images are not segmented and preprocessed, it will cause
errors in the subsequent evaluation of rice appearance quality. Therefore, this paper proposes an
improved instance segmentation network named RiceInstNet based on Mask R-CNN for image
segmentation of adhesive rice backbone network is composed of two improved
VoVNetV2 in parallel, which not only reduces the network parameters on a large scale, but also
strengthens the feature extraction of the adhesive rice grain image. In addition, a sub-network for
learning the object boundary is added on the mask branch, which uses the boundary features to
enrich the mask features and promote the mask prediction to be more experimental
results show that the RiceInstNet can achieve the instance segmentation of adhesive rice g