|
|
Decreased visibility and image defect detection for vehicle mounted camera
Sedláček, Miloš ; Řičánek, Dominik (referee) ; Svědiroh, Stanislav (advisor)
This bachelor thesis deals with the topic of decreased visibility and image defects detection caused by adverse weather conditions or lighting from a vehicle mounted camera. The thesis describes the basic characteristics of the most common influences and their effects on camera data and presents some existing methods of detecting these influences. Next, a dataset containing selected defects is created and described. Afterwards, the issue of artificial neural networks is described in the thesis. A convolutional neural network is implemented for defect detection, which is trained and tested using the dataset. At the end, the achieved results of the network, its computational complexity and comparison with the results of other works are presented.
|
|
|
Recognition of Weather in an Outdoor Stationary Camera View
Jenčo, Michal ; Juránková, Markéta (referee) ; Herout, Adam (advisor)
This thesis deals with classification of weather from stationary outdoor camera images with a landscape view. It classifies fog, clear, partly cloudy and overcast weather in particular. The problem was solved by computing five image flags and using machine learning. Hit rate of 95% was achieved with only small variances between weather types. The main finding of this work is that it is possible to successfully differentiate between the selected weather types with the chosen set of simple image flags. The implementation of this system enables plotting a graph of weather progression during a chosen day.
|
|
|
Multi-Class Weather Classification From Single Images With Convolutional Neural Networks On Embedded Hardware
Bravenec, Tomáš
The paper is focused on creating a lightweight machine learning solution for classificationof weather conditions from input images, that can process the input data in real time on embeddeddevices. The approach to the classification uses deep convolutional neural networks architecture withfocus on lightweight design and fast inference, while providing high accuracy results. The focus oncreating lightweight convolutional neural network architecture capable of classification of weatherconditions also enables usage of the network in real time applications at the edge.
|
|
|
Recognition of Weather in an Outdoor Stationary Camera View
Jenčo, Michal ; Juránková, Markéta (referee) ; Herout, Adam (advisor)
This thesis deals with classification of weather from stationary outdoor camera images with a landscape view. It classifies fog, clear, partly cloudy and overcast weather in particular. The problem was solved by computing five image flags and using machine learning. Hit rate of 95% was achieved with only small variances between weather types. The main finding of this work is that it is possible to successfully differentiate between the selected weather types with the chosen set of simple image flags. The implementation of this system enables plotting a graph of weather progression during a chosen day.
|
guest ::
