Refine
Document Type
Has Fulltext
- no (36)
Is part of the Bibliography
- no (36)
Keywords
- Fluorescence lifetime imaging (9)
- FD-FLIM (8)
- Fluorescence (8)
- Phase shift (8)
- fluorescence (8)
- Neural networks (7)
- Diode lasers (5)
- Fluoreszenz (4)
- Microscopy (4)
- Image processing algorithms (3)
Institute
Holz ist eine der wichtigsten erneuerbaren Ressourcen und kann recycelt werden, wobei derzeit nur unbehandeltes Holz in mehreren Lebenszyklen verwendet werden kann. Allerdings werden in Altholzsortieranlagen meist behandeltes und unbehandeltes Altholz gemischt angeliefert und es gibt keine automatische Klassifizierung zur sortenreinen Sortierung. Frühere Studien zeigen, dass behandelte und unbehandelte Holzproben mit Hilfe von bildgebender Fluoreszenzabklingzeit Mikroskopie unterschieden werden können. Um den Prozess der Klassifizierung von Altholz anhand von Fluoreszenzabklingzeitbildern zu automatisieren, werden zwei neuronale Netzwerke (NN) evaluiert. Das erste NN ist ein MLP (Multi Layer Perceptron), das das Altholz anhand von Erwartungswert und Standardabweichung eines Fluoreszenzabklingzeitbildes klassifiziert. Das zweite untersuchte NN ist ein CNN (Convolutional Neural Network), das das Holz direkt anhand der Fluoreszenzabklingzeitbilder klassifiziert. Beide Netzwerke sind in Python implementiert, um die spätere Verwendung in einem automatischen Klassifizierungs- und Sortierprozess zu erleichtern. Die Evaluation der NN zeigt vielversprechende Ergebnisse zur automatisierten Klassifikation von Altholz.
Development of a Neural Network for Automatic Classification of Post-Consumer Wood Using Rapid-FLIM
(2023)
The economic use of wood is a growing sector, not only because of the significant advantage of wood to retain CO 2 . It is crucial to increase the material recycling of wood in several lifecycles, but currently, there is no reliable post-consumer wood sorting technique in line. This research measures 365 post-consumer wood samples of classes A1-4 four times with the frequency domain fluorescence lifetime imaging microscopy (FD-FLIM) subset method rapid fluorescence lifetime imaging microscopy (Rapid-FLIM). The data is analyzed on their statistical features. Four neural networks based on Multilayer perceptron are then trained and tested with twelve statistical features extracted from the Rapid-FLIM images. The best model for this application contains the optimizer RMSprop, the activation function SELU and the loss function binary crossentropy. The best model of this structure could achieve a false positive ratio of 4.79 % over the ten folds.
The fast classification and identification of plastics presents a significant challenge. The study assesses the suitability of a Multilayer Perceptron to classify and identify commonly found plastic types using Rapid-FLIM, achieving an accuracy of 88.33%. @inproceedingsSchwarz:23, Increasing demand for wood products requires methods to determine its harvest origin and ensure sustainable and legal sourcing. In 15 out of 21 cases, the origin of Quercus robur was differentiable in FD-FLIM studies.
The FD-FLIM technique has a high potential for automated post-consumer wood sorting. A problem of analysing post-consumer wood fluorescence properties is the uncertainty of the post-consumer wood category as the origin of the samples are unknown. In this study, the fluorescence properties of actual post-consumer wood is compared with prepared wood samples. The post-consumer wood samples display slightly different fluorescence intensities and lifetimes due to environmental influences and a higher sample diversity. For improved training of evaluation algorithms for post-consumer wood sorting, the prepared sample set should be extended or the post-consumer wood should be additionally analysed in the laboratory.
Previous research on identifying plastic types and differentiating plastics from environmental material is promising by utilizing the specific fluorescence lifetime, but the evaluation still has to be automated. Therefore, an automated Gaussian analysis is developed for evaluating frequency-domain fluorescence lifetime images of plastics and environmental materials. Furthermore, we applied a “Multilayer Perceptron” and “Random Forest Classifier” to the data resulting from the Gaussian analysis of the frequency domain fluorescence lifetime imaging microscopy data. The classification results show high F1-scores, whereby the best “Multilayer Perceptron” and “Random Forest Classifier” achieved an F1-score of 90%. Thus, identifying and differentiating plastics and environmental materials is possible by applying a “Multilayer Perceptron” or “Random Forest Classifier” to the Gaussian-analyzed imaged fluorescence lifetime data.
Plastic waste is one of the biggest growing factors contributing to environmental pollution. So far there has been no established method to detect and identify plastics in environmental matrices. Thus, a method based on their characteristic fluorescence behavior is used to investigate whether plastics can be detected and identified in tap water under laboratory conditions. The experiments show that the identification of plastics as a function of water depth is possible. As the identification becomes more difficult with higher water depths, investigations with a highly sensitive imaging method were carried out to obtain an areal integration of the fluorescent light and thus better results.
A Novel Approach to Identify Wood Species Optically using Fluorescence Lifetime Imaging Microscopy
(2021)
This contribution presents the results obtained with fluorescence lifetime imaging microscopy (FLIM)within the optical identification and differentiation of the four wood species walnut, beech, spruce, and maple. The experimental setup as well as the evaluation algorithm, with which the experiments were carried out, is explained briefly.
Die wohl größte aktuelle Umweltverschmutzung mit Kunststoffen findet in Ozeanen, Seen und Flüssen statt. Bis jetzt gibt es kein etabliertes Verfahren, mit dem Kunststoffe in Umwelt- bzw. Wassermatrizen detektiert werden können. Deshalb wird mit einem Verfahren, welches auf dem Prinzip der Fluoreszenz basiert, untersucht, ob unter Laborbedingungen Kunststoffe in Wasser detektiert und identifiziert werden können. Anhand eines mathematischen Modells, einer Simulation und mit Hilfe von optischen Experimenten wird gezeigt, dass der Nachweis und auch eine Identifikation von Kunststoffen in Abhängigkeit der Wassertiefe, mit diesem Verfahren möglich ist.
Nowadays, fluorescence lifetime measurements and the method of fluorescence lifetime imaging are already state of the art in biomedical research. Also first investigations show that polymers could be identified due to their endogenous fluorescent lifetime. Thus an approach of identifying polymers with fluorescence lifetime imaging is done in this contribution. Therefore, four different polymers are examined and evaluated with statistical methods in order to determine their specific fluorescence lifetimes.
A novel approach to optically distinguish plastics based on fluorescence lifetime measurements
(2020)
In medical and biological research, fluorescence lifetime measurements and fluorescence lifetime imaging is already a part of the standardized analysis procedures. As first investigations have shown, polymers can be identified using fluorescence lifetime imaging and an evaluation algorithm. Thus, this contribution pursues a novel approach for the direct differentiation of four polymers with fluorescence lifetime imaging. Therefore, the evaluation algorithm is extended to compare several fluorescence lifetime images to prove that a distinction is possible.
The fluorescence decay time allows to identify and distinguish polymers from each other. Three differently treated biaxially-oriented polyethylene terephthalate films are examined with two excitation wavelengths of 445 and 488nm. The fluorescence decay time is dependent of the treatment method of the films and is a means for identification.
Holz ist ein vielseitig einsetzbarer nachwachsender Rohstoff. Neben seinem wirtschaftlichen Nutzen ist er für den Erhalt des Klimas unersetzlich. Eine sortenreine Sortierung für die Weiterverarbeitung von Altholz spielt für einen ressourcenschonenden Umgang eine wichtige Rolle. Um das Potenzial eines neuronalen Netzwerks basierend auf Messdaten der bildgebenden Fluoreszenzabklingzeitmessung für die Altholzsortierung aufzuzeigen, wurden zwei unterschiedliche Klassifikationsansätze auf Basis der Programmiersprache Python gewählt. Die Ergebnisse zeigen, dass die bildbasierte Klassifizierung der Holzart mit einer Genauigkeit von 47,36 % noch ausbaufähig ist. Eine datenbasierte Klassifizierung der Holzart mit einer Identifikationsgenauigkeit von 98,28 % ist dagegen vielversprechend.
A novel approach to optically distinguish plastics based on fluorescence lifetime measurements
(2020)
Today, the automatic separation of polymers from each other in an industrial scale is an unsolved problem. In laboratory environments, two methods are known whereby plastic is sorted either by color or by fluorescence decay time measurements that require fast synchronization and thus expensive equipment. A simple and pragmatic process is proposed to separate plastics from each other: all fluorescent photons are counted in relation to the absorbed photons. A theoretical model and an experimental setup are built in order to determine an apparatus specific quantum efficiency.
Development of a neural network to identify plastics using Fluorescence Lifetime Imaging Microscopy
(2022)
Plastics have become a major part of human’s daily life. An uncontrolled usage of plastic leads to an accumulation in the environment posing a threat to flora and fauna, if not recycled correctly. The correct sorting and recycling of the most commonly available plastic types and an identification of plastic in the environment are important. Fluorescence lifetime imaging microscopy shows a high potential in sorting and identifying plastic types. A data-based and an image-based classification are investigated using python programming language to demonstrate the potential of a neural network based on fluorescence lifetime images to identify plastic types. The results indicate that the data-based classification has a higher identification accuracy compared to the image-based classification.