road lane warning systems, using concepts of computer vision with modules such as OpenCV library and deep learning -- 2

Hello,

I’m building a road lane warning systems, using concepts of computer vision with modules such as OpenCV library and deep learning, a system that uses Machine learning to monitor lane markings and detect when a vehicle is moving out of its lane, a warning system (such as an audio, visual, vibration, or other alert) will notify the motorist of the unintended lane change so the motorist would be able to manoeuvre the vehicle back into its initial lane.

Steps done so far :

Research

Project Proposal

Data set selection and preparation

I have generated my own dataset (977 images)

Data pre-pre-processing

I have pre-processed my data (resizing all the images using OpenCV 184 x 184 size ).

Data Augmentation

This project initially had the original dataset in the ratio of 100:19 as positive to negative class ratio. To balance them appropriately and not generate heavily duplicated data, only 19% of the positive class images have been augmented with Snow cover, Rain cover, Cloud cover and Fog. While, all the augmentations have been applied to all the images in the negative class images. After augmentation all the images have been resized to 184 x 184 size images to keep the image matrix for the deep learning small enough without losing the resolution of the lane markings. After the process was done it resulted in the dataset ratio of 1:1.

Steps pending :

Model Development

This must include an introduction on the tools and infrastructure selected (cloud, Tensorflow etc.). This section should be a maximum of 1000 words plus visualizations and tables. This also must discuss the development of the model, from initial investigations to the final refined version

o Tool/infrastructure selection and rationale (could be multiple):

▪ Azure DSVM, AWS, WEKA

▪ Tensorflow, Scikit-learn, Tableau, PowerBI

o Model Development:

▪ ML algorithm selection

▪ ANN topology

▪ Selection criteria (10FCV, Loss, Accuracy, Sensitivity, Gridsearch)

Performance and Model Outcome

The results should present the results of the experiment/tool. The results should also include techniques used to validate the model/show that it would generalize. The selection of these techniques should also be detailed (with numerical values included). Usually, the findings should also be visualized and/or presented in a tabular manner. The performance should be briefly discussed here. This section should be a maximum of 1000 words plus visualizations and tables.

Because of the pressure, complexity, and lack of time I am finding it difficult to complete this project by myself.

Therefore, I am looking for assistance and willing to pay for it.

I look forward to hearing from you.

Evner: Machine Learning (ML), Kunstig intelligens, Neural Networks

Se mere: expert systems using php, sample systems using data report vb6 mysql, project systems using php, projects using computer vision dspic, payroll systems using aspnet application, microcontroller embedded systems using assembly solutions, inventory systems using vbnet mssql, warning failed move file install modules, warning failed move file joomla modules, sending email using microsoft access modules, cricket score board management systems using oracle, online transaction systems using php mysql, human resourcemanagement systems using aspnet, gps warning systems, invoice systems using php samples, using labview vision, python for computer vision with opencv and deep learning, python for computer vision with opencv and deep learning free download, machine vision library deep learning

Projekt ID: #29475182