Target detection of underground personnel based on deep convolutional neural network
-
摘要: 针对以人为中心的井下视频监控模式存在持续时间受限、多场景同时监视困难、人工监视结果处理不及时等问题，提出了基于深度卷积神经网络的井下人员目标检测方法。首先将输入图片缩放为固定尺寸，通过深度卷积神经网络操作后形成特征图；然后，通过区域建议网络在特征图上形成建议区域，并将建议区域池化为统一大小，送入全连接层进行运算；最后，根据概率分数高低选择最好的建议区域，自动生成需要的目标检测框。测试结果表明，该方法可以成功检测出矿井工作人员的头部目标，准确率达到87.6%。Abstract: In view of problems that human—centered video monitoring mode had limited duration, multiple scenes were difficult to monitor at the same time, and results of manual monitoring were not processed in time, target detection method of underground personnel based on deep convolutional neural network was proposed. Firstly, input image was scaled to a fixed size, and a feature map was formed after operation of deep convolutional neural network; then, a suggestion area was formed on the feature map through area suggestion network, the suggestion area was pooled into a unified size which was sent to full connection layer for operation; finally, the best suggestion area was selected according to probability score, and the required target detection box was automatically generated. The test results show that the method can successfully detect head of underground personnel with an accuracy rate of 87.6%.
-
-
期刊类型引用(7)
1. 任在民，彭飞，刘展志，李志青. 青海省抽水蓄能电站与梯级水库储能电站对比分析. 人民黄河. 2022(S2): 195-198 . 
百度学术
2. 刘天，蒋成成，魏云冰. 风-光-抽水蓄能联合发电系统模型的优化研究. 电工技术. 2021(18): 63-65 . 百度学术
3. 王明松. 风-光-蓄-火联合发电系统的两阶段优化调度策略. 电网与清洁能源. 2020(05): 75-82 . 百度学术
4. 陈希章. 建立基于多能互补的化工园区能源供应体系浅析(二). 化学工业. 2019(03): 42-51 . 百度学术
5. 沈琛云，王明俭，李晓明. 基于风-光-蓄-火联合发电系统的多目标优化调度. 电网与清洁能源. 2019(11): 74-82 . 百度学术
6. 阿旺多杰，嘎玛. 一种适用于偏远山区的多能互补发电系统仿真分析. 电气工程学报. 2019(04): 60-65 . 百度学术
7. 李玉博，贾燕冰. 双馈风机控制系统参数优化. 煤炭技术. 2018(09): 282-285 . 百度学术
其他类型引用(3)
计量
- 文章访问数: 78
- HTML全文浏览量: 13
- PDF下载量: 23
- 被引次数: 10
下载:
