摘要
烧结环节中链篦机台车的侧板由于受到挤压或因温差较大工作环境而产生的裂缝影响球团烧结质量,严重时影响生产.分别采用千眼狼5F04高速相机与imx600y相机拍摄获取链篦机台车侧板裂缝图像并进行标注.使用基于Deeplab V3+的神经网络模型对两种拍摄的侧板裂缝图像数据训练,并对训练数据实验结果对比分析.实验结果表明,使用千眼狼5F04拍摄的图像数据训练的MIoU指标比imx600y高出9.61%,台车侧板裂缝检测也是千眼狼5F04效果更好.为链篦机侧板裂缝的精准检测识别使用摄像机型号提供参考依据.
In the sintering process,the cracks in the side plate of the grate trolley due to extrusion or large temperature difference will affect the sintering quality of pellet and even affect the production seriously.The thousand-eye wolf 5F04 high-speed camera and imx600y camera were used to capture the cracks in the side plate of the grate trolley and mark them.The neural network model based on Deeplab V3+was used to train the two kinds of side slab crack image data taken,and the experimental results of the training data were compared and analyzed.It is shown that the MioU index trained with the image data taken by the Thousand-Eyed Wolf 5F04 is 9.61%higher than that by the imx600y,and the Thousand-Eyed Wolf 5F04 also performs better in detecting cracks in the side plate of the trolley,which may provide a reference for the precise detection and identification of the camera model used in the side plate cracks of the grate machine.
作者
王月明
圣园园
黄文鑫
张建东
吴永刚
WANG Yueming;SHENG Yuanyuan;HUANG Wenxin;ZHANG Jiandong;WU Yonggang(Information Engineering School, Inner Mongolia University of science and technology, Baotou 014010;Sintering Plant, Baotou Iron and Steel (Group) Co., Baotou 014010, China)
出处
《内蒙古科技大学学报》
CAS
2021年第3期282-286,共5页
Journal of Inner Mongolia University of Science and Technology
基金
内蒙古自治区自然基金资助项目(2020MS06008,2019MS06036).
