Task:For the video, rather than sending the actual video, I'm just going to send you the SRT file. That's basically like a subtitle file, that has the coordinates of the drone at all the different times in the video.
Solution ConvertSRT_CSV.py
-
input: videos/DJI_0301.SRT
-
Output: SRT_converted.csv
This program takes the SRT files as input, uses regular expression and gives a CSV file SRT_converted.csv which looks like this:
| TimeStamp | Longitude | Latitude |
|---|---|---|
| 00:00:00,400 | 73.001330 | 19.15014 |
| 00:00:00,500 | 73.001330 | 19.15017 |
| 00:00:00,600 | 73.001323 | 19.15021 |
| ............ | ......... | ........ |
Task: For every second in the video, I want a list of all the images that lie within 35 metres of the drone position.
Solution MainTask_Solution.py
-
Input: SRT_converted.csv, images/
-
Output: resultsDf.csv
-
Time Complexity: O(n.m) n=images, m= video lat long pairs
This program takes the converted SRT file and the images as input and outputs a CSV file. I've used vincenty Formulae to calculate the distance between the latitude, longitude pair even though it takes twice as many mathemetical calculations as the Great circle distance formula because it gives more precise results and since we are dealing within meters, precision is an important factor.
The following images have no GPS data and have been handled(Error log.txt):
1.image DJI_0061.jpg
2.image DJI_0377.jpg
3.image DJI_0452.jpg
4.image DJI_0605.jpg
5.image DJI_0061.jpg
6.image DJI_0377.jpg
7.image DJI_0452.jpg
8.image DJI_0605.jpg
The output CSV looks like this:
| Time | Images |
|---|---|
| 0 | DJI_0010.JPG,DJI_0013.JPG.... |
| 1 | DJI_0013.JPG,DJI_0014.JPG.... |
| 2 | DJI_0057.JPG,DJI_0048.JPG.... |
| . | ............................. |
Task: Along with that, I'll also give you an excel of some points of interest for the client. I want all images within 50 meters of these POIs.
Solution POI_assests_solution.py
-
Input: assets.csv, images/
-
Output: POI_results.csv
This program works similar to the main solution and outputs the desired images to a CSV file.
Task: Also, if possible, give me a KML of the drone path from the video.
Solution Create_KML.py
-
Input: SRT_converted.csv
-
Output: dronePath.kml
A simple program which creates a kml path from the SRT file.
The flight parameters are dynamic and can be adjusted during runtime and any number of images can be handled.