search

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.73k stars 7.96k forks source link

Automate video/picture detection #1173

Open axzxc1236 opened 6 years ago

axzxc1236 commented 6 years ago

First problem is about detecting internet pictures. I tried this command ./darknet detector test data/coco.data cfg/yolov3-tiny-coco.cfg coco.weight http://(IP address)/identification/identification.jpg -i 0 but it doesn't work, I can build a bash script to workaround this problem, but it would be nice if it supports detecting internet picture.

The second problem is how to skip "prediction screen", after using these commands ./darknet detector test data/coco.data cfg/yolov3-tiny-coco.cfg coco.weight identification.jpg -i 0 -ext_output > output.log prediction screen shows up and requires me to close it, is there a command line argument to workaround this problem? (Don't shows the screen, end the program after it completes detection.)

The third problem is can I output timestamp to log? I use ./darknet detector demo data/coco.data cfg/yolov3-tiny-coco.cfg coco.weight video.mp4 -i 0 > output.log and in output.log it shows

FPS:14.3 Objects: motorbike: 44% person: 32% person: 29%

But I don't know where in the video file these objects shows up, is there a way to put something like "time hh:mm:ss" in the log?

AlexeyAB commented 6 years ago

to skip "prediction screen", use flag -dont_show

./darknet detector test data/coco.data cfg/yolov3-tiny-coco.cfg coco.weight identification.jpg -i 0 -ext_output -dont_show > output.log

or

./darknet detector test cfg/coco.data yolov3.cfg yolov3.weights -dont_show -ext_output < data/train.txt > result.txt