Introduction
This project uses a HuskyLens 2 AI camera with a telephoto lens. Through precise focusing, it delivers stable and reliable recognition results for bird targets at different distances.
What You’ll Need
Hardware Preparation
Note: The telephoto lens module will not be available in the DFRobot online store, but it can be purchased through offline channels.
An adjustable telephoto lens is recommended for this project to clearly capture distant birds and achieve better recognition.(The original lens can be replaced by removing the screws on the left and right sides of the lens.)Diagram is shown below:
Install the telephoto lens.
Check HUSKYLENS 2 firmware version: System Settings → Device Information.
If the firmware version is outdated, update it to System Version 1.2.1 or above. For firmware updates, check out this guide:
https://wiki.dfrobot.com/_SKU_SEN0638_Gravity_HUSKYLENS_2_AI_Camera_Vision_Sensor#7.1%20%20Latest%20Version
Software Preparation
Ensure Mind+ V2 is downloaded and installed.
How It Works
Step 1: Make Hardware Connections
Step 2: Train a Bird recognition Model
Method 1: Custom Object Detection Model
Refer to the link below to train an object detection model capable of recognizing bird species and deploy it to the HUSKYLENS 2.
https://community.dfrobot.com/makelog-318320.html
Method 2: Object Tracking Mode
Quick train a bird recognition model using the Object Tracking mode.
1.Select Object Tracking Mode
2.Tap screen to track
3.Set name
4.Recognition
Step 3: Build your bird-watching Camera
3D print and assemble the camera housing.
Step 4: Test
Attachment
The model in the attachment has only been trained on the following bird model:
Attachment.zip 
