针对新学习策略的弱小目标检测方法

Detection Method of Weak and Small Targets for New Learning Strategies

基于深度卷积神经网络的目标检测在应用中展现出了良好的性能,然而,将其应用于弱小目标检测上依然性能欠佳;为了提高弱小目标检测速度和精度的性能,文章提出一种有效的弱小目标检测方法,使用改进的特征提取方法,利用尺度匹配策略选取合适的尺度进行小目标检测;同时在神经网络中设计自适应的融合模块,通过融合特征与接收域以增强目标环境特征;提出的方法在检测速度和精度上都具备良好的性能;有效解决了一般的框策略无法准确获取小目标的问题,新的策略使用自适应参数确定检测框;特别针对海天背景下,检测海面船只问题,提出基于海天线智能分割的方法,进而进行背景处理检测的预处理方法;可以很大程度消除虚警,提高目标检测概率;实验结果表明,提出的方法在视频数据中能够有效检测弱小目标,优于其它先进的目标检测方法。

Based on deep convolution neural network,target detection has a good performance in application,however,its performance is still poor for weak and small target detection.In order to improve the speed and accuracy of the small and weak target detection,an effective small and weak target detection method is proposed,the improved feature extraction method and scale matching strategy are used to select appropriate scale for the small target detection,at the same time,an adaptive fusion module is designed in the neural network,the characteristics of the target environment are enhanced by fusing the features and receiving the domain,the proposed method has a good performance in detection speed and accuracy,and effectively solves the problem that the general box strategy cannot accurately obtain small targets,the new strategy uses adaptive parameters to determine the detection box.Especially for the detection of ships in the sea and sky background,an intelligent segmentation method of the sea and sky line is proposed,and then the background detection preprocessing method carried out,which largely eliminates false alarms and improves the probability of target detection.Experimental results show that the proposed method can effectively detect weak and small targets from the video data,and it is superior to other advanced target detection methods.