Design and Application of Vehicle Mounted Industrial CT Nondestructive Testing System and Underwater Unmanned Equipment Defect Recognition Software
-
摘要: 针对存放地域偏远、不宜频繁运输、难以集中检测的水下无人装备检测需求, 文中基于电源分系统、X射线源分系统、探测采集传输分系统等硬件设备搭建了车载工业计算机断层成像(CT)无损缺陷检测及识别系统, 并结合深度神经网络研制了工业CT检测产品缺陷图像识别软件, 实现对检测图像处理及产品内部缺陷的识别和标识。经实际检测测试, 该系统能够很好地检测和识别水下无人装备内部高能填充物的孔洞、裂隙、脱粘等缺陷, 成像准确率超过98%, 缺陷识别漏检率和虚警率均不超过5%。该系统能够适应产品的不同材料和外形结构, 能够在多种环境下工作, 具有较好的应用前景。Abstract: To satisfy the inspection demands of unmanned undersea equipment that are stored in remote areas, unsuitable for frequent transportation and difficult to conduct centralized testing, this paper builds a vehicle-mounted industrial computed tomography(CT) nondestructive defect detection and identification system with hardware subsystems including power supply, X-ray source, detection, acquisition and transmission. Combined with deep neural networks, an image recognition software for defects in industrial CT inspection products is developed to process inspection images and realize the identification and marking of internal product defects. Practical test results demonstrate that the proposed system can effectively detect and identify pores, cracks, debonding and other internal defects in high-energy fillers of unmanned undersea equipment. The imaging accuracy exceeds 98%, and both the missed detection rate and false alarm rate of defect identification are no more than 5%. This system can adapt to different materials and shapes of the products, and can operate in various environments. It has a promising application prospect.
-
图 1 车载工业CT无损检测系统原理图
Figure 1. Schematic diagram of the vehicle-mounted industrial CT nondestructive testing system
图 2 车载工业CT无损检测系统组成设备
Figure 2. Equipment composition of vehicle-mounted industrial CT nondestructive testing system
图 11 CSPDarknet53骨干网络结构图
Figure 11. Structure diagram of CSPDarknet53 backbone network
图 13 输出特征通道与预测锚框对应关系
Figure 13. Mapping relationship between output feature channels and prediction anchor boxes
图 14 缺陷检测边界框坐标回归计算原理
Figure 14. Calculation principle of coordinate regression for defect detection bounding boxes
图 16 缺陷识别软件检测结果可视化界面
Figure 16. Visual interface of detection results of defect recognition software
表 1 X射线源分系统主要技术参数
Table 1. Key technical parameters of X-ray source subsystem
参数 设置 射线管类型 双焦点双极定向闭合式金属陶瓷射线管 最大管电压 450 kV 额定功率 700 W/ 1500 W焦点规格 0.4 mm/1.0 mm双焦点(符合EN12543标准) 管电流 最大电压下分别为1.6 mA、3.3 mA 电压控制 40~450 kV, 步进0.1 kV, 精确度±0.01% 电流控制 大焦点0.5~3.3 mA, 步进0.1 mA, 精确度±0.01% 射线辐射角 40°×30° 高压发生器
最大功率4.5 kW 冷却方式 油冷 冷却器制冷
功率4500 W
下载: 导出CSV
表 2 成像板主要技术参数
Table 2. Main technical parameters of imaging panel
参数 设置 接收器类型 非晶硅TFT 有效成像尺寸 427 mm×427 mm 像素数量 3 072×3 072 单像素尺寸 139 μm 模数转换位数 16 bit 最高耐压 450 kV 传输接口 千兆以太网 工作温度 10~40 ℃ 供电规格 220VAC(±10%), 50 Hz
下载: 导出CSV
表 3 车载工业CT无损检测系统各部件质量
Table 3. Mass of each component of vehicle-mounted industrial CT nondestructive testing system
序号 部件 质量/t 1 铅房 14.94 2 钢结构与钢板 3.30 3 机械结构 1.30 4 操作间设备 1.00 5 方舱 3.80 6 牵引车 6.80 7 半挂车 3.90 合计 35.04
下载: 导出CSV
表 4 缺陷检测结果
Table 4. Results of defect detection
缺陷类型 mAP@0.5/% 训练集 验证集 测试集 裂隙 91.3 94.5 98.2 孔洞 93.8 96.2 99.1 脱毡 88.6 91.8 97.5 疏松 89.2 93.0 98.7
下载: 导出CSV
表 5 软件性能测试指标
Table 5. Test indicators of software performance
项目 设计性能要求 实测结果 图像像元分辨率 0.1 mm 0.1 mm 最小可识别缺陷尺寸 0.5 mm×0.5 mm 0.5 mm×0.5 mm 支持缺陷类型 4种 裂隙、孔洞、脱毡、疏松 单图检测时间 <5 s 4 096×4 096分辨率图像检测用时3 s
下载: 导出CSV
-
[1] 哈海荣, 王团盟, 鲁忠宝, 等. 一种战斗部用DNAN基炸药热塑态装药改进工艺[J]. 水下无人系统学报, 2020, 28(2): 209-213, 230.Ha H R, Wang T M, Lu Z B, et al. Improved thermoplastic charge technique of DNAN-based explosive for warhead[J]. Journal of Unmanned Undersea Systems, 2020, 28(2): 209-213, 230. [2] 鲁忠宝, 黎勤, 马军利, 等. 鱼雷战斗部装药特点与发展[J]. 水下无人系统学报, 2018, 26(1): 10-15, 45.Lu Z B, Li Q, Ma J L, et al. Research on charge in torpedo warhead, journal of unmanned undersea systems[J]. 2018, 26(1): 10-15, 4. [3] 崔宝成. 浅析医学影像技术学-CT[J]. 世界最新医学信息文摘, 2015, 15(72): 111-112. [4] Xiao Y S, Chen Z Q, Yu D W, et al. The applications of industrial CT NDT technology in geological research[C]//proceedings of the 19th World Conference on Non-Destructive Testing, 2016. [5] 唐庆明, 张明. 固体火箭发动机的装药缺陷及检测方法[J]. 飞航导弹, 2006(7): 52-54. doi: 10.3969/j.issn.1009-1300.2010.01.016 [6] 汤戈, 赵欣雨, 王宇翔, 等. 工业CT技术在地球科学中的应用[J]. CT理论与应用研究, 2024, 33(1): 119-134.Tang G, Zhao X Y, Wang Y X, et al. Applications of industrial computed tomography technology in the geosciences[J]. CT Theory and Applications, 2024, 33(1): 119-134. [7] Ewert U, Fuchs T. Progress in digital industrial radiology. Pt. 2, Computed tomography(CT)[J]. Badania nieniszczące idiagnostyka, 2017(1-2): 7-14. [8] Barciewicz M, Ryniewicz A. The application of computed tomography in the automotive world–how industrial CT works[J]. Technical Transactions, 2018(115): 181-188. doi: 10.4467/2353737xct.18.141.8980 [9] 宁亚倩. 基于X射线CT成像技术的芒果内部缺陷判别与品质检测研究[D]. 武汉: 武汉轻工大学, 2024. [10] De Chiffre L, Carmignato S, Kruth J P, et al. Industrial applications of computed tomography[J]. CIRP annals, 2014, 63(2): 655-677. doi: 10.1016/j.cirp.2014.05.011 [11] Hu S, Xu J K, Lv M C, et al. The Application of Industrial CT Detection Technology in Defects inspection of lithium Ion Battery[J]. Journal of Physics: Conference Series, 2021, 2083(3): 32075. doi: 10.1088/1742-6596/2083/3/032075 [12] 唐盛明, 齐子诚, 刘子瑜, 等. 电子束焊缝超声波C扫描与工业CT检测方法测试结果比较[J]. 无损检测, 2014, 36(10): 49-52,83.Tang S M, Qi Z C, Liu Z Y, at al. Comparison of test method between ultrasonic C-scan and industrial CT in electron beam welding testing[J]. Nondestructive Testing, 2014, 36(10): 49-52,83. [13] 唐盛明, 齐子诚, 郑颖, 等. 工业CT在混凝土钢筋腐蚀检测中的应用[J]. 无损检测, 2017, 39(12): 10-14.Tang S M, Qi Z C, Zheng Y, at al. Application of industrial ct in corrosion detection of concrete reinforcement[J]. Nondestructive Testing, 2017, 39(12): 10-14. [14] Gao Y L, Chen X Q, Liu Y P, et al. Effect of warhead wall thickness on charge uniformityin industrial CT detection[J]. Journal of Measurement Science and Instrumentation, 2017, 8(1): 9-16. [15] 吕延达, 陈亦维, 郭洪勤, 等. 工业CT检测技术在5 kN发动机研制中的应用[J]. 宇航材料工艺, 2020, 50(4): 87-91. doi: 10.12044/j.issn.1007-2330.2020.04.017Lü Y D, Chen Y W, Guo H Q, et al. Application of industrial CT detectior technology in defect inspection of 5 kN engine[J]. Aerospace Materials & Technology, 2020, 50(4): 87-91. doi: 10.12044/j.issn.1007-2330.2020.04.017 [16] 温银堂, 张松, 张玉燕, 等. 新型复合材料界面粘接缺陷的CT检测及表征[J]. 中国测试, 2020, 46(1): 12-17.Wen Y T, Zhang S, Zhang Y Y, et al. CT detection and characterization for interface defects of new composite materials[J]. China Measurement & Test, 2020, 46(1): 12-17. [17] 李智, 陈思宇. 采用X射线CT技术评价沥青路面结构内部质量均匀性[J]. 公路交通科技, 2014, 31(12): 6-11.Li Z, Chen S Y. Evaluating homogeneity of interval quality of asphalt pavement structure with X-ray computed tomography[J]. Journal of Highway and Transportation Research and Development, 2014, 31(12): 6-11. [18] 宋婧, 杨泉. 航天火工产品无损检测技术的应用与发展[J]. 现代制造技术与装备, 2020(6): 164-165. [19] 涂旺, 王文强, 陈佳慧, 等. 增材制造小缺陷的显微CT检测[J]. 无损检测, 2024, 46(5): 50-55,61.Tu W, Wang W Q, Chen J H, et al. Micro-CT detection of small defects in additive manufacturing[J]. Nondestructive Testing, 2024, 46(5): 50-55,61. [20] 孟嘉, 肖鹏. 一种金属增材结构中微缺陷的工业CT检测灵敏度验证方法[J]. 宇航材料工艺, 2024, 54(6): 91-97. doi: 10.12044/j.issn.1007-2330.2024.06.013Meng J, Xiao P. A sensitivity verification method of industrial CT detection for micro-defects in metal additive manufacturing structures[J]. Aerospace Materials & Technology, 2024, 54(6): 91-97. doi: 10.12044/j.issn.1007-2330.2024.06.013 [21] Zuo Z P, Sun T. Application of industrial CT technology in additive manufacturing field[C]// 2022 Photonics & Electromagnetics Research Symposium(PIERS), 2022: 197-203. [22] Wei Z H, Liu B D, Dong B, et al. A joint reconstruction and segmentation method for limited-angle X-Ray tomography[J]. IEEE Access, 2018, 6: 7780-7791. doi: 10.1109/ACCESS.2018.2800719 [23] Fang Z, Wang T. Novel design of industrial real-time CT system based on sparse-view reconstruction and deep-learning image enhancement[J]. Electronics, 2023, 12(8): 1815. doi: 10.3390/electronics12081815 [24] 阳庆国, 谭伯仲. 高性能锥束工业X射线CT系统研制与应用[J]. 光学精密工程, 2023, 31(6): 804-812. doi: 10.37188/OPE.20233106.0804Yang Q G, Tan B Z. Development and application of high-performance cone-beam industrial X-ray CT system[J]. Optics and Precision Engineering, 2023, 31(6): 804-812. doi: 10.37188/OPE.20233106.0804 [25] 吕宁, 崔庆忠, 黄学义, 等. 基于维纳滤波的装药孔隙工业CT图像恢复方法[J]. 科学技术与工程, 2015, 15(30): 132-138. doi: 10.3969/j.issn.1671-1815.2015.30.025Lü N, Cui Q Z, Huang X Y, et al. An industrial CT image restoration method for explosive cavity based on Wiener filtering[J]. Science Technology and Engineering, 2015, 15(30): 132-138. doi: 10.3969/j.issn.1671-1815.2015.30.025 [26] Singh G, Kaur M, Jindal P K, et al. Convolution neural network(CNN) layers in deep learning: A review[J]. AIP Conference Proceedings, 2024, 3121: 40033. doi: 10.1063/5.0221488 [27] Hu F Y, Wan X J, Shen M F, et al. Survey progress on image instance segmentation methods of deep convolutional neural network[J]. Computer Science, 2022, 49(5): 10-24. [28] 魏龙, 刘乐, 刘吉吉, 等. 基于机器学习的固体火箭发动机无损检测数据智能判读[J]. 国防科技, 2021, 42(4): 69-75. doi: 10.13943/j.issn1671-4547.2021.04.12Wei L, Liu L, Liu J J, et al. Intelligent interpretation of non-destructive testing data for solid rocket engines based on machine learning[J]. National Defence Technology, 2021, 42(4): 69-75. doi: 10.13943/j.issn1671-4547.2021.04.12 [29] 常海涛. 基于Faster R-CNN的工业CT图像缺陷检测研究[D]. 兰州: 兰州交通大学, 2018. [30] 张磊, 陈研, 邱焓, 等. 一种车载X射线工业CT系统[J]. 无损探伤, 2021, 45(1): 32-37.Zhang L, Chen Y, Qiu H, et al. Vehicle-mounted X-ray industrial CT system[J]. Nondestructive Testing Technology, 2021, 45(1): 32-37. [31] 杨建学. 一款车载式检查系统的研究开发[J]. 中国机械, 2023(10): 2-8. [32] 李晓晴. 车载式高精度工业CT结构参数快速获取研究[D]. 北京: 清华大学, 2018. [33] Li M, Zhang Z J, Lei L P, et al. Agricultural greenhouses detection in high-resolution satellite images based on convolutional neural networks: Comparison of faster R-CNN, YOLO v3 and SSD[J]. Sensors, 2020, 20(17): 4938. doi: 10.3390/s20174938 [34] Murat A A, Kiran M S. A comprehensive review on YOLO versions for object detection[J]. Engineering Science and Technology, an International Journal, 2025, 70: 102161. doi: 10.1016/j.jestch.2025.102161 [35] Lv Z L, Zhao Z Q, Xia K W, et al. Steel surface defect detection based on MobileViTv2 and YOLOv8[J]. The Journal of Supercomputing, 2024, 80(13): 18919-18941. doi: 10.1007/s11227-024-06248-w [36] 纪泳丞, 李毅, 陈汉平, 等. 基于改进YOLOv11的轻量化裂缝检测算法[J/OL]. 华南理工大学学报(自然科学版), 2026-2-12. http://link.cnki.net/urlid/44.1251.T.20260211.1341.002. [37] Alam M S, Alam M, Tufail M, et al. TobSet: A new tobacco crop and weeds image dataset and its utilization for vision-based spraying by agricultural robots[J]. Applied Sciences, 2022, 12(3): 1308. doi: 10.3390/app12031308 [38] Ocholla I A, Pellikka P, Karanja F, et al. Livestock detection and counting in Kenyan rangelands using aerial imagery and deep learning techniques[J]. Remote Sensing, 2024, 16(16): 2929. doi: 10.3390/rs16162929 -
下载:
点击查看大图
计量
- 文章访问数: 18
- HTML全文浏览量: 13
- PDF下载量: 7
- 被引次数: 0
