基于滤波网络的多余物状态识别研究  

Research on Remainer State Identification Based on Filtering Network Method

作　　者：王宇宸 孙胜利[1,2] 陈夏宁 陈宝岚 马一骏 WANG Yuchen;SUN Shengli;CHEN Xianing;CHEN Baolan;MA Yijun(Shanghai Institute of Technical Physics,Chinese Academy of Sciences,Shanghai 200083,CHN;Key Laboratory of Intelligent Infrared Perception,Chinese Academy of Sciences,Shanghai 200083,CHN;School of Information Science and Technology,ShanghaiTech University,Shanghai 201210,CHN;University of Chinese Academy of Sciences,Beijing 100049,CHN)

机构地区：[1]中国科学院上海技术物理研究所,上海200083 [2]中国科学院智能红外感知重点实验室,上海200083 [3]上海科技大学信息科学与技术学院,上海201210 [4]中国科学院大学,北京100049

出　　处：《半导体光电》2023年第3期471-477,共7页Semiconductor Optoelectronics

基　　金：科技部重点领域创新人才推进计划项目(2019RA4018)。

摘　　要：在航天产品的研发制造过程中,多余物控制至关重要,尤其是多余物状态识别的环节,其核心在于如何有效提取高噪声图像中的局部特征。现有方法尚未针对多余物场景进行有效建模,通用的视觉模型往往会对噪声进行过度拟合,而无法有效滤除噪声信号。为此,提出了一种可学习滤波感知机。该感知机通过采用一种可学习滤波器来替代繁琐的自注意力机制,用以学习空间位置的交互作用信息。随后,引入了频谱掩模用于频域分量特征的抽取,以学习不同频段内的重点信息。实验结果表明,该方法在多余物识别中取得了96.7%的准确率,优于基于卷积和自注意力的模型,并且具有更低的计算复杂度。The remainder control is crucial to the development and manufacturing of aerospace products,and the remainders state recognition is an important part of it.The key which is to effectively extract local features in high noise pictures.However,existing methods have not been modeled well specifically for remainder scenes,and generic vision models are prone to overfitting the noise,making it difficult to filter the noisy signals effectively.To solve this problem,this paper proposes a learnable Filter Network,which replaces the heavy self-attention mechanism by a learnable filter which is used to learn spatial location interaction information.And then incorporates a mask for frequency domain component feature extraction to learn the emphasis information of different frequency bands.It is experimentally demonstrated that this method works better in remainder recognition scenarios,outperforms the convolution and self-attention models,and has better time complexity.

关 键 词：机器视觉 智能制造 多余物控制 航天产品 滤波算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]