AI catching wildlife poachers
Opportunity
Motion detection cameras transmit images to the Wildlife Protection Solutions head office in real time where they are assessed by a team member for evidence of people, vehicles or poaching. This is a slow and labour intensive activity, prompting WPS to look for a solution to streamline the process.
Solution
Silverpond used its tool HighLighter to train a deep learning model to recognise suspected poachers quickly by analysing video streams.
Results
- Doubled detection rate (40% to 80%).
- Seamless integration into existing system.
- Two poachers captured and detained within first week of deployment.
- Over 1 million inferences in the first year alone.
- A CUSTOM AI MODEL WAS developed to account for the types of images captured by WPS' cameras in the field
- The AI model can identify objects of interest without being distracted by other components of the image
- The AI model also detects various animals such as Elephants, giraffes and lions
Detect Objects
Measure Them
Predict Outcomes
Act Decide

2 poachers captured in 1st week

Doubled the detection rate
