提出一种基于Faster RCNN(Faster Region with Convolutional Neural Networks)的电路板缺陷图像自动检测方法。该方法首先应用ResNet50网络作为主干网络以提取缺陷图像特征;然后针对电路板图像中缺陷的极端长宽比特点,提出基于特征金...提出一种基于Faster RCNN(Faster Region with Convolutional Neural Networks)的电路板缺陷图像自动检测方法。该方法首先应用ResNet50网络作为主干网络以提取缺陷图像特征;然后针对电路板图像中缺陷的极端长宽比特点,提出基于特征金字塔的区域推荐网络,得到多尺度特征融合图并生成建议框;最后,通过对感兴趣区域进行池化处理,并结合后续网络实现对电路板图像上缺陷的快速检测。试验证明,所提算法能够快速定位出电路板图像中的各种缺陷,并能实现准确的自动分类识别。展开更多
目的建立气相色谱(GC)测定四川姜黄片中14种农残含量的分析方法。方法以α-BHC、五氯硝基苯、β-BHC、γ-BHC、δ-BHC、三氯杀螨醇、4,4-DDE、4,4-DDD、2,4-DDT、4,4-DDT、氯菊酯、氯氰菊酯、氰戊菊酯、溴氯菊酯为标准物质,Inert Cap 1...目的建立气相色谱(GC)测定四川姜黄片中14种农残含量的分析方法。方法以α-BHC、五氯硝基苯、β-BHC、γ-BHC、δ-BHC、三氯杀螨醇、4,4-DDE、4,4-DDD、2,4-DDT、4,4-DDT、氯菊酯、氯氰菊酯、氰戊菊酯、溴氯菊酯为标准物质,Inert Cap 17为分离柱,以初始温度60℃,30℃/min升至180℃,再以5℃/min升至220℃,2℃/min升至250℃,5℃/min升至270℃,保持35 min为程序升温。四川姜黄片经粉碎,0.1%乙酸乙腈溶液提取,氮气浓缩至近干,正己烷稀释定容,浓硫酸磺化的条件下以外标法定量。结果农残色谱峰的分离度均1.5,理论塔板数均≥10 000,加标回收率70.3%~118.5%,农残检出限0.078~2.22μg/kg。结论该方法前处理操作过程简单,结果重现性好,灵密度高,分离度好,回收率符合要求,适用于四川姜黄片中14种农残含量的测定。展开更多
The functionally graded cemented tungsten carbide (FGCC) is a suitable material choice for cutting tool applications due to balanced hardness and fracture toughness.The presence of cobalt and CaF2 composition gradient...The functionally graded cemented tungsten carbide (FGCC) is a suitable material choice for cutting tool applications due to balanced hardness and fracture toughness.The presence of cobalt and CaF2 composition gradient in FGCC may enhance mechanical as well as antifriction properties.Therefore,structural design of selflubricating FGCC was proposed using Power law composition gradient model and thermal residual stresses (TRSs) as a key parameter.Wherein,S.Suresh and A.Mortensen model was adopted for estimation of TRS,and optimum composition gradient was identified at Power law exponent n =2.The designed material displayed compressive and tensile TRS at surface and core respectively;subsequently fabricated by spark plasma sintering and characterized via scanning electron microscope (SEM),indentation method.The agreement between experimental and analytical values of TRS demonstrated the effectiveness of intended design model in the composition optimization of self-lubricating FGCC.This work will be helpful in implementation of dry machining for clean and green manufacturing.展开更多
文摘提出一种基于Faster RCNN(Faster Region with Convolutional Neural Networks)的电路板缺陷图像自动检测方法。该方法首先应用ResNet50网络作为主干网络以提取缺陷图像特征;然后针对电路板图像中缺陷的极端长宽比特点,提出基于特征金字塔的区域推荐网络,得到多尺度特征融合图并生成建议框;最后,通过对感兴趣区域进行池化处理,并结合后续网络实现对电路板图像上缺陷的快速检测。试验证明,所提算法能够快速定位出电路板图像中的各种缺陷,并能实现准确的自动分类识别。
文摘目的建立气相色谱(GC)测定四川姜黄片中14种农残含量的分析方法。方法以α-BHC、五氯硝基苯、β-BHC、γ-BHC、δ-BHC、三氯杀螨醇、4,4-DDE、4,4-DDD、2,4-DDT、4,4-DDT、氯菊酯、氯氰菊酯、氰戊菊酯、溴氯菊酯为标准物质,Inert Cap 17为分离柱,以初始温度60℃,30℃/min升至180℃,再以5℃/min升至220℃,2℃/min升至250℃,5℃/min升至270℃,保持35 min为程序升温。四川姜黄片经粉碎,0.1%乙酸乙腈溶液提取,氮气浓缩至近干,正己烷稀释定容,浓硫酸磺化的条件下以外标法定量。结果农残色谱峰的分离度均1.5,理论塔板数均≥10 000,加标回收率70.3%~118.5%,农残检出限0.078~2.22μg/kg。结论该方法前处理操作过程简单,结果重现性好,灵密度高,分离度好,回收率符合要求,适用于四川姜黄片中14种农残含量的测定。
基金supported by the Chhattisgarh 12.Council of Science and Technology(CCOST)(Grant No.2230/CCOST/MRP/15).
文摘The functionally graded cemented tungsten carbide (FGCC) is a suitable material choice for cutting tool applications due to balanced hardness and fracture toughness.The presence of cobalt and CaF2 composition gradient in FGCC may enhance mechanical as well as antifriction properties.Therefore,structural design of selflubricating FGCC was proposed using Power law composition gradient model and thermal residual stresses (TRSs) as a key parameter.Wherein,S.Suresh and A.Mortensen model was adopted for estimation of TRS,and optimum composition gradient was identified at Power law exponent n =2.The designed material displayed compressive and tensile TRS at surface and core respectively;subsequently fabricated by spark plasma sintering and characterized via scanning electron microscope (SEM),indentation method.The agreement between experimental and analytical values of TRS demonstrated the effectiveness of intended design model in the composition optimization of self-lubricating FGCC.This work will be helpful in implementation of dry machining for clean and green manufacturing.
基金Project supported by the National Natural Science Foundation of China(No.31271848)the Foundation of Fuli Institute of Food Science of Zhejiang University(No.KY201404),China