Zugl.: Karlsruhe, Univ., Diss., 2004

Universiẗat, Diss., 2002--Heidelberg.

Zugl.: Karlsruhe, Univ., Diss., 2005

Universiẗat, Diss., 2005--Karlsruhe.

Zugl.: Zürich, Techn. Hochsch., Diss., 2006

Depth sensing is pivotal in robotics; however, monocular depth estimation encounters significant challenges. Existing algorithms relying on large-scale labeled data and large Deep Convolutional Neural Networks (DCNNs) hinder real-world applications. We propose two lightweight architectures that achieve commendable accuracy rates of 91.2% and 90.1%, simultaneously reducing the Root Mean Square Error (RMSE) of depth to 4.815 and 5.036. Our lightweight depth model operates at 29-44 FPS on the Jetson Nano GPU, showcasing efficient performance with minimal power consumption. Moreover, we introduce a mask network designed to visualize and analyze the compact depth network, aiding in discerning informative samples for the active learning approach. This contributes to increased model accuracy and enhanced generalization capabilities. Furthermore, our methodology encompasses the introduction of an active learning framework strategically designed to enhance model performance and accuracy by efficiently utilizing limited labeled training data. This novel framework outperforms previous studies by achieving commendable results with only 18.3% utilization of the KITTI Odometry dataset. This performance reflects a skillful balance between computational efficiency and accuracy, tailored for low-cost devices while reducing data training requirements.:1. Introduction 2. Literature Review 3. AI Technologies for Edge Computing 4. Monocular Depth Estimation Methodology 5. Implementation 6. Result and Evaluation 7. Conclusion and Future Scope Appendix

This dissertation describes the improvement of signal analysis in ultrasonic microscopy for nondestructive testing. Specimens with many thin layers, like modern electronic components, pose a particular challenge for identifying and localizing defects. In this thesis, new evaluation algorithms have been developed which enable analysis of highly complex layer-stacks. This is achieved by a specific evaluation of multiple reflections, a newly developed iterative reconstruction and deconvolution algorithm, and the use of classification algorithms with a highly optimized simulation algorithm. Deep delaminations inside a 19-layer component can now not only be detected, but also localized. The new analysis methods also enable precise determination of elastic material parameters, sound velocities, thicknesses, and densities of multiple layers. The highly improved precision of determined reflections parameters with deconvolution also provides better and more conclusive results with common analysis methods. / Die vorgelegte Dissertation befasst sich mit der Verbesserung der Signalauswertung für die Ultraschallmikroskopie in der zerstörungsfreien Prüfung. Insbesondere bei Proben mit vielen dünnen Schichten, wie bei modernen Halbleiterbauelementen, ist das Auffinden und die Bestimmung der Lage von Fehlstellen eine große Herausforderung. In dieser Arbeit wurden neue Auswertealgorithmen entwickelt, die eine Analyse hochkomplexer Schichtabfolgen ermöglichen. Erreicht wird dies durch die gezielte Auswertung von Mehrfachreflexionen, einen neu entwickelten iterativen Rekonstruktions- und Entfaltungsalgorithmus und die Nutzung von Klassifikationsalgorithmen im Zusammenspiel mit einem hoch optimierten neu entwickelten Simulationsalgorithmus. Dadurch ist es erstmals möglich, tief liegende Delaminationen in einem 19-schichtigem Halbleiterbauelement nicht nur zu detektieren, sondern auch zu lokalisieren. Die neuen Analysemethoden ermöglichen des Weiteren eine genaue Bestimmung von elastischen Materialparametern, Schallgeschwindigkeiten, Dicken und Dichten mehrschichtiger Proben. Durch die stark verbesserte Genauigkeit der Reflexionsparameterbestimmung mittels Signalentfaltung lassen sich auch mit klassischen Analysemethoden deutlich bessere und aussagekräftigere Ergebnisse erzielen. Aus den Erkenntnissen dieser Dissertation wurde ein Ultraschall-Analyseprogramm entwickelt, das diese komplexen Funktionen auf einer gut bedienbaren Oberfläche bereitstellt und bereits praktisch genutzt wird.

This dissertation describes the improvement of signal analysis in ultrasonic microscopy for nondestructive testing. Specimens with many thin layers, like modern electronic components, pose a particular challenge for identifying and localizing defects. In this thesis, new evaluation algorithms have been developed which enable analysis of highly complex layer-stacks. This is achieved by a specific evaluation of multiple reflections, a newly developed iterative reconstruction and deconvolution algorithm, and the use of classification algorithms with a highly optimized simulation algorithm. Deep delaminations inside a 19-layer component can now not only be detected, but also localized. The new analysis methods also enable precise determination of elastic material parameters, sound velocities, thicknesses, and densities of multiple layers. The highly improved precision of determined reflections parameters with deconvolution also provides better and more conclusive results with common analysis methods.:Kurzfassung......................................................................................................................II Abstract.............................................................................................................................V List ob abbreviations........................................................................................................X 1 Introduction.......................................................................................................................1 1.1 Motivation.....................................................................................................................2 1.2 System theoretical description.....................................................................................3 1.3 Structure of the thesis..................................................................................................6 2 Sound field.........................................................................................................................8 2.1 Sound field measurement............................................................................................8 2.2 Sound field modeling..................................................................................................11 2.2.1 Reflection and transmission coefficients.........................................................11 2.2.2 Sound field modeling with plane waves..........................................................13 2.2.3 Generalized sound field position.....................................................................19 2.3 Receiving transducer signal.......................................................................................20 2.3.1 Calculation of the transducer signal from the sound field...............................20 2.3.2 Received signal amplitude..............................................................................21 2.3.3 Measurement of reference signals..................................................................24 3 Ultrasonic Simulation......................................................................................................27 3.1 State of the art............................................................................................................27 3.2 Simulation approach..................................................................................................28 3.2.1 Sound field measurement based simulation...................................................28 3.2.2 Reference signal based simulation.................................................................30 3.3 Determination of the impulse response.....................................................................31 3.3.1 1D ray-trace algorithm....................................................................................31 3.3.2 2D ray-trace algorithm....................................................................................33 3.3.3 Complexity reduction – optimizations.............................................................35 4 Deconvolution – Determination of reflection parameters............................................38 4.1 State of the art............................................................................................................39 4.1.1 Decomposition techniques..............................................................................39 4.1.2 Deconvolution.................................................................................................41 4.2 Analytic signal investigations for deconvolution.........................................................42 4.3 Single reference pulse deconvolution........................................................................44 4.4 Multi-pulse deconvolution..........................................................................................47 4.4.1 Homogeneous multi-pulse deconvolution.......................................................48 4.4.2 Multi-pulse deconvolution with simulated GSP profile....................................49 5 Reconstruction.................................................................................................................50 5.1 State of the art............................................................................................................50 5.2 Reconstruction approach...........................................................................................51 5.3 Direct material parameter estimation.........................................................................52 5.3.1 Sound velocities and layer thickness..............................................................52 5.3.2 Density, elastic modules and acoustic attenuation.........................................54 5.4 Iterative material parameter determination of a single layer......................................56 5.5 Reconstruction of complex specimens......................................................................60 5.5.1 Material characterization of multiple layers ....................................................60 5.5.2 Iterative simulation parameter optimization with correlation...........................62 5.5.3 Pattern recognition reconstruction of specimens with known base structure. 66 6 Applications and results.................................................................................................71 6.1 Analysis of stacked components................................................................................71 6.2 Time-of-flight and material analysis...........................................................................74 7 Conclusions and perspectives.......................................................................................78 References.......................................................................................................................82 Figures.............................................................................................................................86 Tables...............................................................................................................................88 Appendix..........................................................................................................................89 Acknowledgments.........................................................................................................100 Danksagung...................................................................................................................101 / Die vorgelegte Dissertation befasst sich mit der Verbesserung der Signalauswertung für die Ultraschallmikroskopie in der zerstörungsfreien Prüfung. Insbesondere bei Proben mit vielen dünnen Schichten, wie bei modernen Halbleiterbauelementen, ist das Auffinden und die Bestimmung der Lage von Fehlstellen eine große Herausforderung. In dieser Arbeit wurden neue Auswertealgorithmen entwickelt, die eine Analyse hochkomplexer Schichtabfolgen ermöglichen. Erreicht wird dies durch die gezielte Auswertung von Mehrfachreflexionen, einen neu entwickelten iterativen Rekonstruktions- und Entfaltungsalgorithmus und die Nutzung von Klassifikationsalgorithmen im Zusammenspiel mit einem hoch optimierten neu entwickelten Simulationsalgorithmus. Dadurch ist es erstmals möglich, tief liegende Delaminationen in einem 19-schichtigem Halbleiterbauelement nicht nur zu detektieren, sondern auch zu lokalisieren. Die neuen Analysemethoden ermöglichen des Weiteren eine genaue Bestimmung von elastischen Materialparametern, Schallgeschwindigkeiten, Dicken und Dichten mehrschichtiger Proben. Durch die stark verbesserte Genauigkeit der Reflexionsparameterbestimmung mittels Signalentfaltung lassen sich auch mit klassischen Analysemethoden deutlich bessere und aussagekräftigere Ergebnisse erzielen. Aus den Erkenntnissen dieser Dissertation wurde ein Ultraschall-Analyseprogramm entwickelt, das diese komplexen Funktionen auf einer gut bedienbaren Oberfläche bereitstellt und bereits praktisch genutzt wird.:Kurzfassung......................................................................................................................II Abstract.............................................................................................................................V List ob abbreviations........................................................................................................X 1 Introduction.......................................................................................................................1 1.1 Motivation.....................................................................................................................2 1.2 System theoretical description.....................................................................................3 1.3 Structure of the thesis..................................................................................................6 2 Sound field.........................................................................................................................8 2.1 Sound field measurement............................................................................................8 2.2 Sound field modeling..................................................................................................11 2.2.1 Reflection and transmission coefficients.........................................................11 2.2.2 Sound field modeling with plane waves..........................................................13 2.2.3 Generalized sound field position.....................................................................19 2.3 Receiving transducer signal.......................................................................................20 2.3.1 Calculation of the transducer signal from the sound field...............................20 2.3.2 Received signal amplitude..............................................................................21 2.3.3 Measurement of reference signals..................................................................24 3 Ultrasonic Simulation......................................................................................................27 3.1 State of the art............................................................................................................27 3.2 Simulation approach..................................................................................................28 3.2.1 Sound field measurement based simulation...................................................28 3.2.2 Reference signal based simulation.................................................................30 3.3 Determination of the impulse response.....................................................................31 3.3.1 1D ray-trace algorithm....................................................................................31 3.3.2 2D ray-trace algorithm....................................................................................33 3.3.3 Complexity reduction – optimizations.............................................................35 4 Deconvolution – Determination of reflection parameters............................................38 4.1 State of the art............................................................................................................39 4.1.1 Decomposition techniques..............................................................................39 4.1.2 Deconvolution.................................................................................................41 4.2 Analytic signal investigations for deconvolution.........................................................42 4.3 Single reference pulse deconvolution........................................................................44 4.4 Multi-pulse deconvolution..........................................................................................47 4.4.1 Homogeneous multi-pulse deconvolution.......................................................48 4.4.2 Multi-pulse deconvolution with simulated GSP profile....................................49 5 Reconstruction.................................................................................................................50 5.1 State of the art............................................................................................................50 5.2 Reconstruction approach...........................................................................................51 5.3 Direct material parameter estimation.........................................................................52 5.3.1 Sound velocities and layer thickness..............................................................52 5.3.2 Density, elastic modules and acoustic attenuation.........................................54 5.4 Iterative material parameter determination of a single layer......................................56 5.5 Reconstruction of complex specimens......................................................................60 5.5.1 Material characterization of multiple layers ....................................................60 5.5.2 Iterative simulation parameter optimization with correlation...........................62 5.5.3 Pattern recognition reconstruction of specimens with known base structure. 66 6 Applications and results.................................................................................................71 6.1 Analysis of stacked components................................................................................71 6.2 Time-of-flight and material analysis...........................................................................74 7 Conclusions and perspectives.......................................................................................78 References.......................................................................................................................82 Figures.............................................................................................................................86 Tables...............................................................................................................................88 Appendix..........................................................................................................................89 Acknowledgments.........................................................................................................100 Danksagung...................................................................................................................101