共查询到19条相似文献,搜索用时 390 毫秒
1.
使用遗传算法,对从炭素制品缺陷特征中提取的参数进行优化处理,然后对优化后的参数进行模糊分类。鉴于得到的优化参数空间尺度分布大且数量级不一,通过建立数学模型,首先对参数在空间尺度进行处理,使参数分布更加集中,然后计算各样本间的相似度,进而确定分类的阈值,最后通过决策算法对炭素制品X射线图像中的缺陷进行分类,从而实现对炭素制品缺陷的分类识别。 相似文献
2.
3.
炭素制品缺陷的X射线自动检测技术研究 总被引:4,自引:3,他引:1
针对炭素制品X光图像的特点,对其缺陷的提取与识别技术进行了研究,给出了目标边界提取算法和基于小波变换的图像增强算法,实现了图像的背景去除及增强处理。在此基础上,为排除噪声干扰的影响,采用数学形态学和迭代阈值分割相结合的方法从背景去除后的图像中提取出缺陷区域,取得了良好的效果。对缺陷特征选择及识别方法进行了研究,设计了基于遗传策略的特征选择和基于BP神经网络的缺陷识别算法,计算表明:缺陷正确识别率可达95%以上。采用上述技术开发完成了一套炭素制品缺陷X射线自动检测系统。 相似文献
4.
对埋弧焊X射线焊缝圆形和线形缺陷图像进行分析,针对焊缝缺陷局部图像强噪声、弱对比度和常规方法不易区分类型的特点,将主成成分分析的思想引入焊缝圆形和线形缺陷类型分类。分析缺陷疑似局部图像自相关矩阵特征值发现,圆形线形焊缝缺陷疑似局部图像分类问题可降维为一维问题,极大地简化计算和提高运算速度。基于此给出圆形和线性缺陷分类算法,由于将缺陷图像分类问题降维,使得分类算法对模板的选择具有较强的鲁棒性。通过现场超过400张焊缝缺陷局部图像的实验表明,无论如何选取模板,线性缺陷的识别率均在98%以上,圆形缺陷的识别率在89%~98.8%之间,且在16次模板更换实验中,4次圆形缺陷识别率达到98.8%。 相似文献
5.
《仪器仪表学报》2020,(3)
针对传统的铸件缺陷检测不能对缺陷进行分类分级等问题,提出了一种基于Mask R-CNN的铸件X射线DR图像缺陷检测算法。首先对原始图像进行预处理,采用引导滤波进行图像平滑,平滑图像与原图像进行差分得到差分图像,将差分图像与平滑图像相加运算使图像增强,再利用Labelme进行图像标注,形成训练数据集。送入Mask R-CNN深度学习网络,通过特征提取网络生成建议区域,分类、回归网络生成边界框和掩码,经多次参数调节后得到训练网络模型,最后测试数据集。实验数据结果表明,气泡1~5级的检测率分别为:66.7%,71.4%,77.4%,88.9%,87.5%;疏松1~5级检测率为:62.5%,72.2%,77.1%,83.3%,81.1%。检测结果证明应用Mask R-CNN结合引导滤波增强方法的缺陷检测方法可以较好的实现对铸件X射线DR图像的缺陷检测的分级分类,为工业铸件缺陷检测提供了应用深度学习方法的解决方案。 相似文献
6.
7.
摘要:针对传统的铸件缺陷检测不能对缺陷进行分类分级等问题,提出了一种基于Mask R CNN的铸件X射线DR图像缺陷检测算法。首先对原始图像进行预处理,采用引导滤波进行图像平滑,平滑图像与原图像进行差分得到差分图像,将差分图像与平滑图像相加运算使图像增强,再利用Labelme进行图像标注,形成训练数据集。送入Mask R CNN深度学习网络,通过特征提取网络生成建议区域,分类、回归网络生成边界框和掩码,经多次参数调节后得到训练网络模型,最后测试数据集。实验数据结果表明,气泡1~5级的检测率分别为:667%,714%,774%,889%,875%;疏松1~5级检测率为:625%,722%,771%,833%,811%。检测结果证明应用Mask R CNN结合引导滤波增强方法的缺陷检测方法可以较好的实现对铸件X射线DR图像的缺陷检测的分级分类,为工业铸件缺陷检测提供了应用深度学习方法的解决方案。 .txt 相似文献
8.
9.
航空发动机叶片实时成像自动检测技术研究 总被引:5,自引:1,他引:5
为了实现航空发动机叶片X射线实时成像自动检测,以基于平板探测器的X射线实时成像系统为研究对象,根据射线实时成像的特点,对航空发动机叶片缺陷的提取技术进行了研究。根据航空发动机叶片X射线数字图像灰度变化的特点,将叶片图像划分为6个区域,分别进行基于扫描线的自适应中值滤波模拟出缺陷的背景:将原始图像与背景图像相减,获得背景平坦、缺陷突出的差值图像;通过对差值图像进行阈值分割,实现对缺陷的分离;用数学形态学的开运算对缺陷图像滤波以去除噪声;为保证缺陷的准确性,用区域生长法重新生长出缺陷:最后提取出缺陷的特征参数。将缺陷提取的结果和人工评片的结果比较,证明这种方法是准确、有效的,为实现航空发动机叶片X射线自动检测打下了良好的基础。 相似文献
10.
11.
赵元黎 《仪表技术与传感器》2003,(7):43-44
应用光声效应检测粉煤灰的碳含量,因制备样品存在的差异和原煤成分不同等原因,即使在碳含量相同时,检测到的光声信号也有一定的差异。采用模糊分类方法可进行正确分类。对两种模糊分类方法进行了研究,指出模糊分类是设计光声效应检测粉煤灰中碳含量的一种有效方法。并用样品分类实验说明了这种方法的可行性及实用性。模糊分类器是一软件分类器,不增加仪器硬件开销,是设计光声光谱仪器的一种好方法。 相似文献
12.
《Mechanical Systems and Signal Processing》2007,21(1):441-456
Today the knowledge of a process is very important for engineers to find optimal combination of control parameters warranting productivity, quality and functioning without defects and failures. In our laboratory, we carry out research in the field of high speed machining with modelling, simulation and experimental approaches. The aim of our investigation is to develop a software allowing the cutting conditions optimisation to limit the number of predictive tests, and the process monitoring to prevent any trouble during machining operations. This software is based on models and experimental data sets which constitute the knowledge of the process. In this paper, we deal with the problem of vibrations occurring during a machining operation. These vibrations may cause some failures and defects to the process, like workpiece surface alteration and rapid tool wear. To measure on line the tool micro-movements, we equipped a lathe with a specific instrumentation using eddy current sensors. Obtained signals were correlated with surface finish and a signal processing algorithm was used to determine if a test is stable or unstable. Then, a fuzzy classification method was proposed to classify the tests in a space defined by the width of cut and the cutting speed. Finally, it was shown that the fuzzy classification takes into account of the measurements incertitude to compute the stability limit or stability lobes of the process. 相似文献
13.
B. T. Budai B. T. Porodnov I. V. Myakutina 《Russian Journal of Nondestructive Testing》2014,50(5):279-282
High-precision and environmentally friendly measurement of sheet product parameters is an important problem. Well-known environmentally friendly high-precision optical triangulation meters for measuring the parameters of sheet products enable one to measure the parameters only for longitudinal defects in sheet products. It is shown here that the modernization of a high-precision optical triangulation meter designed for measuring longitudinal defects in sheet metal allows trans-verse defects to be precisely measured in sheet products as well. Such a meter allows one to control the main rolling parameters in addition to measuring the longitudinal defects in sheet products. This provides environmentally friendly production of high-grade sheet products. 相似文献
14.
15.
塑料注射成型工艺参数优化的模糊规则网络模型 总被引:1,自引:0,他引:1
注射成型是塑料产品成型的最主要工艺,工艺参数是影响成型产品外观、尺寸与性能的关键因素之一。工艺参数的设置与优化属于弱理论、强经验的问题,迫切需要发展科学化、系统化的方法。针对产品缺陷修正中人工经验依赖性强的问题,构建知识的统一模糊化规则形式,建立工艺优化知识表示和推理于一体的Takagi-Sugeno-Kang(TSK)模糊规则网络模型。进一步,提出从工艺数据集自动发现工艺参数优化规则的学习方法,采用Dropout策略与Bagging集成学习策略缓解高维工艺数据下工艺知识库增长出现的规则数量爆炸等问题。分析了模糊规则网络参数、结构对知识表示和推理的影响,建立模型的参数学习与结构优化的双重进化方法。提出基于经验回放的工艺数据增量学习方法,建立数据的增量学习策略。在注射成型工艺数据集上的结果表明,模型的规则数量和长度降低了50%,具有高可解释性以及增量学习稳定性。 相似文献
16.
17.
Ceramic and tile industries should indispensably include a grading stage to quantify the quality of products. Actually, human control systems are often used for grading purposes. An automatic grading system is essential to enhance the quality control and marketing of the products. Since there generally exist six different types of defects originating from various stages of tile manufacturing lines with distinct textures and morphologies, many image processing techniques have been proposed for defect detection. In this paper, a survey has been made on the pattern recognition and image processing algorithms which have been used to detect surface defects. Each method appears to be limited for detecting some subgroup of defects. The detection techniques may be divided into three main groups: statistical pattern recognition, feature vector extraction and texture/image classification. The methods such as wavelet transform, filtering, morphology and contourlet transform are more effective for pre-processing tasks. Others including statistical methods, neural networks and model-based algorithms can be applied to extract the surface defects. Although, statistical methods are often appropriate for identification of large defects such as Spots, but techniques such as wavelet processing provide an acceptable response for detection of small defects such as Pinhole. A thorough survey is made in this paper on the existing algorithms in each subgroup. Also, the evaluation parameters are discussed including supervised and unsupervised parameters. Using various performance parameters, different defect detection algorithms are compared and evaluated. 相似文献
18.
在无损检测信号处理和特征构造的基础上 ,用神经网络对缺陷进行识别 ,然后运用模糊积分对多个神经网络的分类结果进行融合。并以胶接结构典型缺陷的超声波检测与识别为例 ,给出了一些实验结果。结果表明用模糊积分集成后的正确分类率比单独网络的正确分类率有较大提高 相似文献
19.
将模糊控制和神经网络理论相结合,通过遗传算法对其参数进行优化,有效地解决了常规模糊理论不能自学习和神经网络算法易陷入局部极小、收敛速度慢等缺点,并对其应用于电力变压器故障诊断进行了仿真,实例仿真结果表明该算法具有较快的收敛速度和较高的计算精度,故障诊断结果证实了该算法应用于电力变压器故障诊断的有效性。 相似文献