基于SSD 的小目标特征强化检测算法

为解决原始单次多框目标检测(single shot multibox detector,SSD)目标检测算法中对小目标物体检测能力不足的问题,提出一种改进的SSD目标检测算法.采用VGG19作为特征提取网络,在低层特征图部分引入Conv3_3卷积特征图,对Conv4_4进行转置卷积操作,将转置卷积后得到的Conv4_3同Conv3_3的特征图进行特征拼接,实验部分使用VOC数据集对模型进行训练与测试.结果表明:该算法可提高检测能力,目标检测精度能比原始SSD算法提高3.6％,小目标检测效果比改进前也有明显提升.

Enhanced Detection Algorithm of Small Target Based on SSD

In order to solve the problem of insufficient ability of small target detection in the original single shot multibox detector (SSD) target detection algorithm, an improved SSD target detection algorithm is proposed. VGG19 is used as the feature extraction network, Conv3_3 convolution feature graph is introduced into the low-level feature map, the transpose convolution operation is carried out on Conv4_4, and the Conv4_3 obtained by transpose convolution is spliced with the feature map of Conv3_3. In the test, the VOC data set is used to train and test the model. The results show that the algorithm can improve the detection ability, the target detection accuracy can be improved by 3.6% compared with the original SSD algorithm, and the effect of small target detection is also significantly improved.