[number of seconds to group by]

[YYYYWD]

I don't know what is the mean of " [number of seconds to group by] "

how can i do if i want to python plot.py ?

[creotiv]

that's my bad english) Should be [number of seconds to aggregate by] It's just saying width in seconds of window which will be aggregated in one value.

[YYYYWD]

OK fine. But I find some problems. This is your the past code

I don‘t know how to find these 'car' and 'bus'. Because your the test video of 'report.csv' just has

So I have to change the code

Could you tell me how I should do that?

---

And If I change the Green shade's position ranges.What should I do?

[creotiv]

yes, because stats has only vehicles and not type classification, you need to change plot to use stats that you have. You did it.

Green area is an exit zone, where cars counting happens. To change them, you need to change EXIT_PTS variable in the traffic.py. it just (x,y) points that are creating poligon(it can be more then 4 to create any figure you need.)

[YYYYWD]

OK, thank you very much! If I want to identify other car models, do I still need to train classifiers myself? I have data set of MATLAB, so how can I apply it to it?

[creotiv]

Better use pretrained models first like YOLO3 or SSD, if they work, then just use them.

[YYYYWD]

Yes，my data is from the YOLO. And I have an idea. I want to judge the abnormal behavior of the vehicle.I want to base on the trajectory of the vehicle.If I can use your code's 'path' to do that.Or do you have any better Suggestions?

[YYYYWD]

I have a new question. I used my own test video, which changed the export area. But it was found that only two cars could be detected at any one time. If there are three cars left, they can't be detected. What do I need to change?

[zubairahmed-ai]

Facing the same issue KeyError: 'car'

If you are recommending to use your own video then please also change the code to remove this key and use the generic 'vehicles' key as you suggested above