|
Traffic Violation Detection on Crossroads
Karpíšek, Miroslav ; Bartl, Vojtěch (referee) ; Špaňhel, Jakub (advisor)
This bachelor thesis presents procedure for the detection of red-light violation. In the theoretical part of the thesis, the current solution aproaches used in image processing are described. The practical part focuses on creation of program for automatic traffic corridors detection, vehicle tracking and the current traffic light state detection. The results obtained by experimenting with the proposed procedure and the possibilities of its further improvement are also discussed.
|
|
Detection of Graffiti Tags in Image
Fischer, Martin ; Kodym, Oldřich (referee) ; Špaňhel, Jakub (advisor)
The aim of this work is to compare different approaches of computer vision with the intention of automatic detection of graffiti tags in the image. The solution was based on models based on neural networks. Both the proven detection models and the experimental models were tested here. The most accurate one (Faster R-CNN) achieved an accuracy of 83% mAP, indicating the suitability of these models to the tag detection problem.
|
|
Detection of Vehicle License Plates in Video
Líbal, Tomáš ; Hradiš, Michal (referee) ; Herout, Adam (advisor)
This thesis deals with preparation of training dataset and training of convolutional neural network for licence plate detection in video. Darknet technology was used for detection, specifically the YOLOv3-tiny neural network model. The solution was focused on the most accurate detection and the smallest number of false positives per image, thus minimizing overall model error. Dataset was prepared from existing freely available datasets, from the dataset provided by the GRAPH@FIT research group, and from self-annotated images created from downloaded YouTube videos. Furthermore, this dataset has been processed using data augmentation, extending it to twice the size. The YOLO Mark tool was used to create annotations. An ROC curve was used to visualize the detection success. Created solution reaches minimum total error 10,849%. Part of the solution is already mentioned dataset.
|
|
Exploitation of images from panoramatic camera of mobile mapping system
Belanis, Pavel ; Buday, Michal (referee) ; Volařík, Tomáš (advisor)
This diploma thesis deals with an automated detection of vertical traffic signs in images from the panoramic camera Ladybug5. From the detected signs with help of a classifier, a GIS data set is automatically created, usable for example to passportisation of traffic signs. The first part of the thesis describes a theoretical basis needed to understand the given problematics. The second part is devoted to a specific procedure leading to the reliable classifier, its testing on an independent set of images and automated creation of the GIS data set. The output of the work are the trained classifiers and the GIS data sets containing vertical traffic signs.
|
|
Detection of objects for industrial robots using computer vision
Huber, Michal ; Škrabánek, Pavel (referee) ; Parák, Roman (advisor)
The aim of this bachelor thesis is to create an image processing algorithm based on data captured by camera. Introduction of this thesis deals with the current situation of object detection in industrial applications and briefly presents Raspberry Pi and OpenCV library. Following chapters deal with Raspberry Pi setup, test objects design and captured pictures modifications. Last section of these chapters is devoted to a design of an image processing algorithm. The conclusion of this thesis deals with a design of vizualization interface and also describes a laboratory experiment to test the functionality of a designed algorithm.
|
|
Computer Controlled Model of Vehicle
Deingruber, Ondřej ; Musil, Petr (referee) ; Zemčík, Pavel (advisor)
This bachelor's thesis covers the topic of controlling models using servos and motors. The target was to design and construct a functional model and use it to demonstrate the principles of controlling models by computer. The work was extended by implementing vehicle detection program used for combat of the vehicles. A vehicle moving on tracks, controlled by two DC motors and one servo was constructed. For the control of the vehicle was selected a combination of single-board computer and microcontroller. For the purposes of detecting the vehicle OpenCV library paired with a camera and IR LEDs was used.
|
|
Object detection for video surveillance using the SSD approach
Dobranský, Marek ; Lokoč, Jakub (advisor) ; Božovský, Petr (referee)
The surveillance cameras serve various purposes ranging from security to traffic monitoring and marketing. However, with the increasing quantity of utilized cameras, manual video monitoring has become too laborious. In re- cent years, a lot of development in artificial intelligence has been focused on processing the video data automatically and then outputting the desired no- tifications and statistics. This thesis studies the state-of-the-art deep learning models for object detection in a surveillance video and takes an in-depth look at SSD architecture. We aim to enhance the performance of SSD by updating its underlying feature extraction network. We propose to replace the initially used VGG model by a selection of modern ResNet, Xception and NASNet classifica- tion networks. The experiments show that the ResNet50 model offers the best trade-off between speed and precision, while significantly outperforming VGG. With a series of modifications, we improved the Xception model to match the ResNet performance. On top of the architecture-based improvements, we ana- lyze the relationship between SSD and a number of detected classes and their selection. We also designed and implemented a new detector with the use of temporal context provided by the video frames. This detector delivers enhanced precision while...
|
|
Application of computational methods in classification of glass stones
Lébl, Matěj ; Hnětynková, Iveta (advisor)
Application of computational methods in classification of glass stones Bc. Matěj Lébl Abstrakt: The goal of this thesis is to employ mathematical image processing methods in automatic quality control of glass jewellery stones. The main math- ematical subject is a matrix of specific attributes representing digital image of the studied products. First, the thesis summarizes mathematical definition of digital image and some standard image processing methods. Then, a complete solution to the considered problem is presented. The solution consists of stone localization within the image followed by analysis of the localized area. Two lo- calization approaches are presented. The first is based on the matrix convolution and optimized through the Fourier transform. The second uses mathematical methods of thresholding and median filtering, and data projection into one di- mension. The localized area is analyzed based on statistical distribution of the stone brightness. All methods are implemented in the MATLAB environment. 1
|
|
Visual Localization of an Object in 3D Space
Bátoryová, Jana ; Barták, Roman (advisor) ; Obdržálek, David (referee)
The purpose of this project is to propose and implement a system for object localization using a stereo vision - two cameras. The system computes the position of the cameras relatively to each other using a calibration pattern. Then a user selects an object to track. Different algorithms can be used for tracking. Both detection-based and sequence-based approaches can be used. When the object is found in the view of both cameras, the system estimates a position of the object in three-dimensional space using triangulation and displays the results live.
|
|
Detector of the Human Head in Image
Svoboda, Jakub ; Orság, Filip (referee) ; Goldmann, Tomáš (advisor)
Detection of human head is an important part of person detection and identification algorithms. This thesis is focused on the detection of human head with methods based on neural networks. The majority the of conventional detectors can identify objects within a limited range of positions, whereas models based on neural networks offer a more robust approach. In this thesis we trained the current state-of-the-art models and compared their accuracy and speed. The most accurate model proved to be RetinaNet which has reached 85.15% AP. This detector can be used to improve current available algorithms for person detection, identification and tracking.
|