- Railway track crack detection system project ppt install#
- Railway track crack detection system project ppt software#
Railway track crack detection system project ppt install#
If you install another version, I don't warrant about result.
So you have to install these first.Īlso, I made this project on CUDA 10.0 and cuDNN 7.3 environment. Our project based on anaconda, tensorflow-gpu, jupyter notebook and etc.
Railway track crack detection system project ppt software#
Lee Juhee, Department of Software in Soongsil University Setting for training Training model(You can skip this to download pretrained models)īy: Kim Minseok, Department of Software in Soongsil University /.The original author's Githeub code address is as follows. We used some of EdjeElectronics's code to design the model. Obstacle Detection model : Faster-RCNN-Inception-V2.Patent Registeration - Railroad Obstacle Detection System(RODS) 2019.11.30.Korean Software Registeration - Railway Obstacle Detection System(RODS) / Railroad Tracker 2019.11.30.Bronze Prize on Software Contest In Soongsil Univercity 2019.11.07.And because the number of used data is low, we've use Augmentation it in a variety of ways. Our ultimate goal is to detect and signal the driver, even when any obstacles are detected, but because this project was designed for demonstration rather than for actual commercialization, we planned to learn only a few pre-selected obstacles and demonstrate them in a controlled environment. The data used for learning was prepared with track and train models and taken in a controlled environment. Also we selected Object Detection Deep Learning Model to recognize obstacle. For this purpose, the image of the track and the masked data were studied to create a Segmentation Model. The system is designed to minimize casualties and property damage by sending a signal to the engineer when the detected obstacle is on the track and there is a possibility of serious casualties or equipment damage in the event of a collision. Our topic is to install a camera on the front of the train to detect tracks and obstacles. The objective of the project is to improve problem resolution and understanding of the model by using a variety of deep learning models to meet the challenges faced in the field. This project is an ongoing project with the support of Spartan SW project to study Computer Vision of Soongsil University.