Software pro měření defektů závitových hnízd
Název práce v češtině: | Software pro měření defektů závitových hnízd |
---|---|
Název v anglickém jazyce: | Development of Software for Defect Measurement on Threaded Inserts |
Klíčová slova: | fotogrametrie|závitové hnízdo|počítačové vidění|měření defektů|hloubková mapa|přehledový snímek|zpracování videa |
Klíčová slova anglicky: | photogrammetry|threaded insert|computer vision|defect measurement|depth map|one image overview|video processing |
Akademický rok vypsání: | 2024/2025 |
Typ práce: | diplomová práce |
Jazyk práce: | |
Ústav: | Katedra softwaru a výuky informatiky (32-KSVI) |
Vedoucí / školitel: | RNDr. Jan Blažek, Ph.D. |
Řešitel: |
Zásady pro vypracování |
Objective:
The objective of this diploma thesis is to develop software capable of accurately measuring defects on threaded inserts used in power plant vessel flanges. The software will utilize input data consisting of high-resolution 4K+ videoscans of threaded inserts with diameters ranging from 60mm to 120mm. These videos will capture the internal threads by rotating the camera 360° and performing several axial shifts. Additionally, the dataset will be supplemented with training data comprising 3D scans of silicone impressions of the threads. Tasks: 1) Develop algorithms to process the videoscans and generate a one image overview composed of frames from the input video data 2) Implement techniques for generating a high-precision depth map of the threaded inserts registered with one image overview. 3) Evaluate limits of the software by means of input data quality and its parametrization. 4) Document the software development process, including methodologies, algorithms, and implementation details. Deliverables: - Software capable of: - Directly process video stream from GoPro Hero 9 as well as MP4 files - Generating overview images from input videoscans. - Producing high-precision depth maps of threaded inserts. - Detecting and measuring defects with accuracy. - Thesis report documenting the entire development process, including literature review, methodology, results, and conclusions. |
Seznam odborné literatury |
Yang, Lihe, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. "Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data." In CVPR, 2024. [Online]. Available: https://depth-anything.github.io |