Difference between revisions of "Camera detect"
Jump to navigation
Jump to search
Clone the gihub repo - git clone https://github.com/pjreddie/darknet.git then cd darknet then make
Grab the pre-trained weights - wget https://pjreddie.com/media/files/yolov3.weights
Then begin to play - e.g. ./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg
| Line 4: | Line 4: | ||
To ensure that CT stays with us forever, we needed to have a way to automatically count ducks (she previously manually counted 1 million cows for us from individual pictures...and we were worried that she would run away when we told her about counting ducks at Troups Creek). So we found this awesome program called YoloV3 and decided to see if this model (trained on a generically available database). | To ensure that CT stays with us forever, we needed to have a way to automatically count ducks (she previously manually counted 1 million cows for us from individual pictures...and we were worried that she would run away when we told her about counting ducks at Troups Creek). So we found this awesome program called YoloV3 and decided to see if this model (trained on a generically available database). | ||
<br><br> | <br><br> | ||
| − | Steps to recreate this process is as follows (largely following https://pjreddie.com/darknet/yolo/): | + | Steps to recreate this process is as follows on a linux operating system (largely following https://pjreddie.com/darknet/yolo/): |
<li>Clone the gihub repo - git clone https://github.com/pjreddie/darknet.git then cd darknet then make</li> | <li>Clone the gihub repo - git clone https://github.com/pjreddie/darknet.git then cd darknet then make</li> | ||
<li>Grab the pre-trained weights - wget https://pjreddie.com/media/files/yolov3.weights</li> | <li>Grab the pre-trained weights - wget https://pjreddie.com/media/files/yolov3.weights</li> | ||
<li>Then begin to play - e.g. ./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg</li> | <li>Then begin to play - e.g. ./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg</li> | ||
| + | <br> | ||
| + | We now can optimise the system to begin playing further. We have set this as a web interface to allow some playing around. | ||
Revision as of 14:08, 17 March 2023
Outline
This website outlines the work we have been doing to analyse images for objects.
Case study 1 - Troups Creek Duck Counting
To ensure that CT stays with us forever, we needed to have a way to automatically count ducks (she previously manually counted 1 million cows for us from individual pictures...and we were worried that she would run away when we told her about counting ducks at Troups Creek). So we found this awesome program called YoloV3 and decided to see if this model (trained on a generically available database).
Steps to recreate this process is as follows on a linux operating system (largely following https://pjreddie.com/darknet/yolo/):
We now can optimise the system to begin playing further. We have set this as a web interface to allow some playing around.
