Vehware Truck Camera Project 2018 -- Oguzhan Tilki

Week 1

Title: Second Semester Plan

Date: 01/28/2019

Attendance: Oguzhan Tilki

Accomplishments

- Worked on related algorithms

Problems

- Couldn't find a suitable environment to check algorithms

TODO

- Will have a discussion about project future with Merwyn Jones.

Week 2

Title: Company Meeting

Date: 02/03/2019

Attendance: Bill Howe, Oguzhan Tilki , Rana Ece Ertem

Meeting with William Howe

- Bill gave us seperate work to do mine is implementing the alternative algorithm and Ece will implement same thing we did last semester.
- He want me to do it in realtime because new algorithm has a good efficiency when it comes to framing.
- He gave us Ip cameras to try with our computers (to check if they are compatible).
- Our former colleague Swapnil will give company computer back soon to start working on it.
- There was a hardware issue between Ece and Bill (Bill will provide a converter to Ece (Macbook Issues :D) )

Problems

- 1 week will not be enough to implement a brand new idea.

TODO

- I need to discuss project's future with Merwyn Jones.

Week 5

Title: Company Meeting

Date: 02/28/2019

Attendance: Bill Howe, Oguzhan Tilki, Ezgi Durmazpinar, Rana Ece Ertem

Meeting with William Howe

- This week I showed everything that I was working on for 3 weeks.
- The code provides us many options in the perspective of algorithms it shows 7 different algorithms performance (Frames per second).
- It showed there is a performance gap up to tenfold in some cases between Mosse and other 6. 
- At this point, we don't have auto Region of Interest that means we need to give a square shape manually to track it but it is possible to make it auto. 
- I showed the program on some different videos thus we figured out how it works and variances between performances.

Problems

- Manual ROI made some misunderstandings.

TODO

- I need to make auto ROI. 

Week 6

Title: Company Meeting

Date: 03/07/2019

Attendance: Bill Howe, Oguzhan Tilki

Meeting with William Howe

- I showed the result that I get from using different algorithms last week and Bill wants us to integrate it with other code.
- I searched for a solution to integration there were some examples but they are far beyond what we do.
- This week I didn't show anything to Bill because nothing has been implemented in the last 7 days. 
- At this point, I need to work with someone from vehware to get a result faster.

Problems

- I Couldn't find any document about integration yet.

TODO

- Need to show what I achieved to Prof Jones and prepare documents that he wanted from me.

Week 7

Title: Company Meeting

Date: 03/14/2019

Attendance: Bill Howe, Oguzhan Tilki, Ezgi Durmazpinar, Rana Ece Ertem

Meeting with William Howe

- Ezgi and I worked on the integration of MOSSE algorithm with our existing code.
- We made lots of searches and couldn't find any tool or library that could make integration possible.
- Because of their working principles, they can't work parallel and they are alternative to each other. 
- At this point, Bill gave separate tasks to every group member and mine is calculating the center of motion points and showing the coordinates of it.

Problems

- It will take more than a semester with a team to create the same structure that we have with using Mosse and I work on it just myself.

- I achieved a good result but we decide to not continue on that way because of lack of time (But I will give the code to vehware so that next semester new students can pursue and get better results).

TODO

- Find the way to get center of the motion points and its coordinates.

Week 8

Title: Spring Break

Date: 03/16/2019 - 03/23/2019

AIM

- Find the way to get center of the motion points and its coordinates.

PROGRESS

- Found how to display what I got on the same window with actual output to make all data is on one screen.

Week 9

Title: Cancelled Meeting

Date: 03/23/2019 - 03/31/2019

AIM

- Find the way to get center of the motion points and its coordinates.

PROGRESS

- Completely find to how to determine to center of all tracking dots and displaying it with taking advantages of the OpenCV library functions.

Week 10

Title: Vehware Meeting

Date: 04/04/2019

Attendance: Bill Howe, Oguzhan Tilki, Ezgi Durmazpinar, Rana Ece Ertem, and Prof. Jones

ACHIEVEMENT

- I made more than what I expected to do and calculated the centers of dots at the danger and warning zones separately.
- Made a Demo to make sure everybody understands how it works.
- Now we got the center of the moving objects so that we can track them easier.

TO DO

- Bill wanted me to merge the zones and consider them as one so that we will be able to track one car if it is on both danger and warning zone.
- Another thing that wanted is assigning a threshold value for the number of the dots to consider it as a car or an inaccurate tracking dot. 
- After determining if it is a car or not putting a circle to its center.

Week 11

Title: Vehware Meeting

Date: 04/11/2019

Attendance: Bill Howe, Oguzhan Tilki, Ezgi Durmazpinar, Rana Ece Ertem, and Prof. Jones

ACHIEVEMENT

- I have merged the zones as one to find a single center of dots rather than two different centers.
- Still working on deciding the threshold value, it is hard to decide because I need many tests to understand if the value is suitable for deciding cars and I need to do it for every value.
- Discussed frontal application of what we did on side of the truck and it will have some complications but we need to try.

TO DO

- Demo of merged zones with the best possible threshold value and with the circle at the center.
- Bill wanted me to divide the screens and display my result on another screen without irrelevant tracking dots(not sure how to do).

Week 12

Title: Vehware Meeting

Date: 04/18/2019

Attendance: Bill Howe, Oguzhan Tilki, Ezgi Durmazpinar, Rana Ece Ertem

- This week we made the demo of the latest versions of what we have one by one. 
- Bill gave some suggestions about the presentation that we will do next week.

TO DO

- Some little improvements on codes' cosmetic and preparing for presentation next week

Week 13

Title: Final Presentation

Date: 04/25/2019

-  This week we presented everything we have done in this semester.
-  Ezgi and I worked on our codes separately to make them also work on Frontal view.

TO DO

- Combine Ezgi's and my code to one (Ezgis code finds vectors and their magnitude, My code finds the center of motion and display the data on screen).

Week 14

Title: Meeting with Prof Jones

Date: 05/02/2019

Attendance: Prof Jones, Oguzhan Tilki, Ezgi Durmazpinar, Rana Ece Ertem

- We do not have any other meeting with Mr. Howe Presentation was the last Demo for him.
- As expected we merged the codes (vector and center).
- We showed Prof Jones in his Computer how it works.

TO DO

- We need to write README and add some comment to the code.