In foggy weather, images of outdoor scene are usually characterized with poor visibility as well as faint color saturation. The degraded hazy images may have substantial negative impact on most computer vision systems...In foggy weather, images of outdoor scene are usually characterized with poor visibility as well as faint color saturation. The degraded hazy images may have substantial negative impact on most computer vision systems. Thus image haze removal is of the practical significance in engineering. This paper proposes a fast and effective single image haze removal algorithm on the basis of the physics imaging model. To extract the global atmospheric light accurately, we exploit multiple prior rules underlying hazy images, and put forward a novel measurement to judge the likelihood that a pixel is regarded as the global atmospheric light. In addition, the rough transmission map is estimated through a multiscale fusion process based on the Laplace pyramid transform, and refined by a total variation model. Experimental results demonstrate the proposed method outperforms most of the state-of-the-art algorithms in terms of the dehazing quality, and achieves a trade-off between the computational efficiency and haze removal capability.展开更多
Rapid cooling and solidification during laser additive manufacturing(LAM)can produce ultra-fine microstructure with higher strength.However,the non-uniform cell/grain structure can easily result in early stress concen...Rapid cooling and solidification during laser additive manufacturing(LAM)can produce ultra-fine microstructure with higher strength.However,the non-uniform cell/grain structure can easily result in early stress concentration and fracture during deformation,which remains a major challenge for the LAM field.Using Al-12Si as the model alloy,we employed the external static magnetic field(SMF)to modulate the laser powder bed fusion process(L-PBF),demonstrating a uniform microstructure with a refined cell structure.The mechanical properties show that the SMF can produce a combination of high tensile strength of 451.4±0.5 MPa and large uniform elongation of 10.4%±0.79%,which are superior to those of previously-reported Al-Si alloys with post-treatment or element alloying.The mechanism analysis based on multi-scale simulation reveals the determining role of SMF in rapid solidification,and this method is applicable to the microstructure control of other metallic materials during LAM.展开更多
Laser powder bed fusion(LPBF)of high-strength Al alloys is challenging due to the formation of both hot and cold cracks.In the present work,highly dense and crack-free AA2024 samples could be additively manufactured v...Laser powder bed fusion(LPBF)of high-strength Al alloys is challenging due to the formation of both hot and cold cracks.In the present work,highly dense and crack-free AA2024 samples could be additively manufactured via inoculation treatment of Zr-based metallic glass(MG)powders.The columnar grains in the LPBF-fabricated AA2024 alloy were significantly refined and almost completely transformed to the equiaxed grains with a bimodal grain size distribution consisting of ultrafine grains with a size smaller than 1μm and relatively coarser grains.The grain refinement can be associated with the formation of Al3Zr particles,serving as the heterogeneous nucleation sites for theα-Al matrix.Complete routes for columnar-equiaxed-transition(CET)have been revealed by tailoring the concentration of nucleation particles and solidification conditions.CET occurs both at the melt pool boundary due to the sufficiently high concentration of Al3Zr particles and at the topmost of the melt pool due to the heterogeneous nucleation driven by constitutional undercooling.Between these two regions,columnar grains epitaxially grow with orientations determined by the thermal gradient.The as-built Zr-based MG inoculated AA2024 specimens are robust in healing hot cracks due to a more tortuous propagation path of cracks for equiaxed grains.The as-fabricated AA2024/5%MG specimens exhibit a high ultimate tensile strength of 531 MPa due to crack elimination and grain refinement,surpassing most of the reported values for the LPBF-fabricated AA2024 alloy inoculated with other inoculated powders.The present work could provide a novel inoculation agent to fabricate high-strength Al alloys and the CET can be used to precisely control the grain morphology.展开更多
In modern energy systems,substations are the core of electricity transmission and distribution.However,similar appearance and small size pose significant challenges for automatic identification of electrical devices.T...In modern energy systems,substations are the core of electricity transmission and distribution.However,similar appearance and small size pose significant challenges for automatic identification of electrical devices.To address these issues,we collect and annotate the substation rotated device dataset(SRDD).Further,feature fusion and feature refinement network(F3RNet)are constructed based on the classic structure pattern of backbone-neck-head.Considering the similar appearance of electrical devices,the deconvolution fusion module(DFM)is designed to enhance the expression of feature information.The balanced feature pyramid(BFP)is embedded to aggregate the global feature.The feature refinement is constructed to adjust the original feature maps by considering the feature alignment between the anchors and devices.It can generate more accurate feature vectors.To address the problem of sample imbalance between electrical devices,the gradient harmonized mechanism(GHM)loss is utilized to adjust the weight of each sample.The ablation experiments are conducted on the SRDD dataset.F3RNet achieves the best detection performance compared with classical object detection networks.Also,it is verified that the features from global feature maps can effectively recognize the similar and small devices.展开更多
基金supported by the National Natural Science Foundation of China(61571241)the Industry-University-research Prospective Joint Project of Jiangsu Province(BY2014014)+2 种基金the Major Projects of Jiangsu Province University Natural Science Research(15KJA510002)the Jiangsu Province Graduate Research and Innovation Project(CXZZ130476)the Science Research Fund of NUPT(NY215169)
文摘In foggy weather, images of outdoor scene are usually characterized with poor visibility as well as faint color saturation. The degraded hazy images may have substantial negative impact on most computer vision systems. Thus image haze removal is of the practical significance in engineering. This paper proposes a fast and effective single image haze removal algorithm on the basis of the physics imaging model. To extract the global atmospheric light accurately, we exploit multiple prior rules underlying hazy images, and put forward a novel measurement to judge the likelihood that a pixel is regarded as the global atmospheric light. In addition, the rough transmission map is estimated through a multiscale fusion process based on the Laplace pyramid transform, and refined by a total variation model. Experimental results demonstrate the proposed method outperforms most of the state-of-the-art algorithms in terms of the dehazing quality, and achieves a trade-off between the computational efficiency and haze removal capability.
基金the National Key Research and Development Program of China(No.2019YFA0705300,2021YFB3702502)the National Natural Science Foundation of China(Nos.52001191,52127807,52271035)+4 种基金Natural Science Foundation of Shanghai(No.23ZR1421500)SPMI Project from Shanghai Academy of Spaceflight Technology(No.SPMI2022-06)Independent Research Project of State Key Laboratory of Advanced Special Steel,Shanghai Key Laboratory of Advanced FerrometallurgyShanghai University(No.SKLASS 2022-Z10)the Science and Technology Commission of Shanghai Municipality(No.19DZ2270200).
文摘Rapid cooling and solidification during laser additive manufacturing(LAM)can produce ultra-fine microstructure with higher strength.However,the non-uniform cell/grain structure can easily result in early stress concentration and fracture during deformation,which remains a major challenge for the LAM field.Using Al-12Si as the model alloy,we employed the external static magnetic field(SMF)to modulate the laser powder bed fusion process(L-PBF),demonstrating a uniform microstructure with a refined cell structure.The mechanical properties show that the SMF can produce a combination of high tensile strength of 451.4±0.5 MPa and large uniform elongation of 10.4%±0.79%,which are superior to those of previously-reported Al-Si alloys with post-treatment or element alloying.The mechanism analysis based on multi-scale simulation reveals the determining role of SMF in rapid solidification,and this method is applicable to the microstructure control of other metallic materials during LAM.
基金supported by the National Natural Science Foundation of China(Nos.52175330 and 51805267)the Frontier Leading Technology Basic Research Project of Jiangsu(No.BK20202007)the Fundamental Research Funds for the Central Universities(No.30921011202).
文摘Laser powder bed fusion(LPBF)of high-strength Al alloys is challenging due to the formation of both hot and cold cracks.In the present work,highly dense and crack-free AA2024 samples could be additively manufactured via inoculation treatment of Zr-based metallic glass(MG)powders.The columnar grains in the LPBF-fabricated AA2024 alloy were significantly refined and almost completely transformed to the equiaxed grains with a bimodal grain size distribution consisting of ultrafine grains with a size smaller than 1μm and relatively coarser grains.The grain refinement can be associated with the formation of Al3Zr particles,serving as the heterogeneous nucleation sites for theα-Al matrix.Complete routes for columnar-equiaxed-transition(CET)have been revealed by tailoring the concentration of nucleation particles and solidification conditions.CET occurs both at the melt pool boundary due to the sufficiently high concentration of Al3Zr particles and at the topmost of the melt pool due to the heterogeneous nucleation driven by constitutional undercooling.Between these two regions,columnar grains epitaxially grow with orientations determined by the thermal gradient.The as-built Zr-based MG inoculated AA2024 specimens are robust in healing hot cracks due to a more tortuous propagation path of cracks for equiaxed grains.The as-fabricated AA2024/5%MG specimens exhibit a high ultimate tensile strength of 531 MPa due to crack elimination and grain refinement,surpassing most of the reported values for the LPBF-fabricated AA2024 alloy inoculated with other inoculated powders.The present work could provide a novel inoculation agent to fabricate high-strength Al alloys and the CET can be used to precisely control the grain morphology.
基金This work was supported by Science and Technology Project of State Grid Corporation of China(Research and application of audiovisual active perception and collaborative cognitive technology for smart grid operation and maintenance scenarios)(5600–202046347 A-0–0–00).
文摘In modern energy systems,substations are the core of electricity transmission and distribution.However,similar appearance and small size pose significant challenges for automatic identification of electrical devices.To address these issues,we collect and annotate the substation rotated device dataset(SRDD).Further,feature fusion and feature refinement network(F3RNet)are constructed based on the classic structure pattern of backbone-neck-head.Considering the similar appearance of electrical devices,the deconvolution fusion module(DFM)is designed to enhance the expression of feature information.The balanced feature pyramid(BFP)is embedded to aggregate the global feature.The feature refinement is constructed to adjust the original feature maps by considering the feature alignment between the anchors and devices.It can generate more accurate feature vectors.To address the problem of sample imbalance between electrical devices,the gradient harmonized mechanism(GHM)loss is utilized to adjust the weight of each sample.The ablation experiments are conducted on the SRDD dataset.F3RNet achieves the best detection performance compared with classical object detection networks.Also,it is verified that the features from global feature maps can effectively recognize the similar and small devices.
