|
|
Analyse of photovoltaic solar cells by the photoluminescence method
Baura, Tomáš ; Bača, Petr (referee) ; Vaněk, Jiří (advisor)
This thesis deals with the analysis of solar cells by photoluminescence method. Photoluminescence method is based on the excitation of luminescence radiation of the solar cell material by an external excitation light source. This method can detect various types of defects in the material solar cells. The main objective is the design and realization of a measuring system defects of solar cells, which is based on this method. For excitation of luminescence is used LED array. For the detection of luminescence is used CCD camera with IR optical filter. To filter out the excitation light radiation is used disc screen and optical interrupter. The parameters and options of this measuring system are verified and characterized by test measuring for example the achievable rotation speed of screen and response of optical interrupter. At the end is tested detection of luminescence radiation of solar cells in the measuring system. The measured images are compared with the method of electroluminescence.
|
|
|
Thermovision of photovoltaic modules automatic analysis
Repko, Ilia ; Křivík, Petr (referee) ; Vaněk, Jiří (advisor)
This work deals with the automatic evaluation of thermographic images of photovoltaic modules. In theoretical part of work described main principles of sun's battery work and methods for detection defects, that affecting the quality work, including method thermography, which is based on principle of contactless measuring the surface temperature the observed object. Practical part dedicated to creation of algorithms for detection defects, result is a source code for the program MATLAB.
|
|
|
Deep Neural Networks for Defect Detection
Juřica, Tomáš ; Herout, Adam (referee) ; Hradiš, Michal (advisor)
The goal of this work is to bring automatic defect detection to the manufacturing process of plastic cards. A card is considered defective when it is contaminated with a dust particle or a hair. The main challenges I am facing to accomplish this task are a very few training data samples (214 images), small area of target defects in context of an entire card (average defect area is 0.0068 \% of the card) and also very complex background the detection task is performed on. In order to accomplish the task, I decided to use Mask R-CNN detection algorithm combined with augmentation techniques such as synthetic dataset generation. I trained the model on the synthetic dataset consisting of 20 000 images. This way I was able to create a model performing 0.83 AP at 0.1 IoU on the original data test set.
|
|
|
Inovation of system for electroluminiscence defect detection of solar cells
Lepík, Pavel ; Křivík, Petr (referee) ; Vaněk, Jiří (advisor)
This master thesis analyses the existing methods both practically and theoretically used to detect defected surface area in solar cells. Various methods were used but by using an upgraded CMOS camera without IR filter to implement the electroluminescence method, this has proven to have a very crucial impact on the results. Given the overall results and the acquired information, a procedure with a simple parameter can be setup to carry out the measurements. In addition to this a catalog was formed showing the defects occurring in mono and polycrystalline solar cells.
|
|
|
Classification of board defects in semiconductor manufacturing
Jašek, Filip ; Vágner, Martin (referee) ; Dřínovský, Jiří (advisor)
This diploma thesis focuses on detecting defects in semiconductor wafer manufacturing. It explores methods for identifying faulty chips and controlling yield during production. To classify defects machine learning techniques are used. Initially, ResNet18 architecture was used for inference, but low accuracy was attributed to limited input data. Transfer learning with ResNet50v2 was then attempted, resulting in improved metric with different dataset. Hyperparameter tuning and data augmentations were also explored. The study found that autoencoders for data compression during inference increased speed but led to degraded evaluation metrics.
|
|
|
Automation of nuclear fuel visual inspection
Knotek, Jaroslav ; Blažek, Jan (advisor) ; Horáček, Jan (referee)
The safety and performance of nuclear plant relies, among others, on the quality of nuclear fuel. The quality fulfilling designed criteria of the fuel in use is inspected and reported on periodically. Visual inspection focuses on the condition of the fuel based on its visual properties. During the inspection, the fuel is being recorded and analysed by the inspector. The current state of the fuel assemblies is compared to the historical statistics which helps do decide whether this particular assembly remains or gets replaced. This thesis describe a project initiated by Centrum Výzkumu Řež focusing on digital image processing methods application to visual inspection process. The result of the project is a tool that accelerates the process of report making. Firstly, it transforms the inspection video into one image overview and highlight a significant part (more than 95%) of possible defects to the inspector. 1
|
|
|
Detection of infill defects in 3D printed structures using the DIC method
Doležal, Tomáš ; Halabuk, Dávid (referee) ; Ščerba, Bořek (advisor)
Additive manufacturing offers wide range of advantages. However various internal defects are likely to be formed during the manufacturing process which negatively affect mechanical properties. Detection of those defects is critical to ensure that the manufactured component stays reliable and maintains its dependability throughout its whole life. The potential of the digital image correlation (DIC) method for detect detection in components manufactured using additive technologies has not yet been investigated in the literature. In this master thesis a novel non-destructive defectoscopic method for the internal defect detection in 3D printed structures is presented, based on the evaluation of the strain field obtained by the DIC method. The method was experimentally evaluated on samples with artificial internal defects fabricated by FDM technology. The samples containing defects were successfully visually detected. A convolutional neural network was then used for the defect detection and achieved a classification accuracy of 94,5 %. This methodology has a potential to provide cheap and fast detection of internal defects formed in additive manufactured components in the future although future research is still required.
|
|
|
Classification of board defects in semiconductor manufacturing
Jašek, Filip ; Vágner, Martin (referee) ; Dřínovský, Jiří (advisor)
This diploma thesis focuses on detecting defects in semiconductor wafer manufacturing. It explores methods for identifying faulty chips and controlling yield during production. To classify defects machine learning techniques are used. Initially, ResNet18 architecture was used for inference, but low accuracy was attributed to limited input data. Transfer learning with ResNet50v2 was then attempted, resulting in improved metric with different dataset. Hyperparameter tuning and data augmentations were also explored. The study found that autoencoders for data compression during inference increased speed but led to degraded evaluation metrics.
|
|
|
Automation of nuclear fuel visual inspection
Knotek, Jaroslav ; Blažek, Jan (advisor) ; Horáček, Jan (referee)
The safety and performance of nuclear plant relies, among others, on the quality of nuclear fuel. The quality fulfilling designed criteria of the fuel in use is inspected and reported on periodically. Visual inspection focuses on the condition of the fuel based on its visual properties. During the inspection, the fuel is being recorded and analysed by the inspector. The current state of the fuel assemblies is compared to the historical statistics which helps do decide whether this particular assembly remains or gets replaced. This thesis describe a project initiated by Centrum Výzkumu Řež focusing on digital image processing methods application to visual inspection process. The result of the project is a tool that accelerates the process of report making. Firstly, it transforms the inspection video into one image overview and highlight a significant part (more than 95%) of possible defects to the inspector. 1
|
|
|
Deep Neural Networks for Defect Detection
Juřica, Tomáš ; Herout, Adam (referee) ; Hradiš, Michal (advisor)
The goal of this work is to bring automatic defect detection to the manufacturing process of plastic cards. A card is considered defective when it is contaminated with a dust particle or a hair. The main challenges I am facing to accomplish this task are a very few training data samples (214 images), small area of target defects in context of an entire card (average defect area is 0.0068 \% of the card) and also very complex background the detection task is performed on. In order to accomplish the task, I decided to use Mask R-CNN detection algorithm combined with augmentation techniques such as synthetic dataset generation. I trained the model on the synthetic dataset consisting of 20 000 images. This way I was able to create a model performing 0.83 AP at 0.1 IoU on the original data test set.
|
guest ::
