Bird Buddy was crafted with a specific vision: to deliver beautiful bird photos to your fingertips every day. Unlike traditional trail cams, which capture every movement that can result in hundreds of subpar images, Bird Buddy's essence lies in curating exceptional moments.
In this article, you can read more about:
The Postcard-creation process
1. Detection: A bird has to be near the center of the perch to trigger the laser motion detector. This precise positioning ensures that our camera captures the bird in an optimal pose, setting the stage for a great photo. A word of caution: While we encourage creativity, note that DIY or non-Bird Buddy add-ons can impact the proximity of the birds. Since our camera is calibrated specifically for Bird Buddy-designed systems, if birds perch too far, the camera might not get triggered.
2. Recording & AI Assessment: Once our motion detector springs into action the camera will start capturing both photos and videos. What happens next is a blend of cutting-edge technology and our passion for birds. Our own bird-recognizing AI model takes the reins. This intelligent model meticulously scans each image, hunting for clear, well-framed snapshots and identifying the bird species captured.
But here's the thing: our AI is selective. If during a session all photos are found blurry or if they don't present the bird in the best possible light—say, the bird is too close for an aesthetically pleasing shot—the model will tactfully discard these less-than-ideal photos.
Behind the Scenes: Our AI Model in Action
Our AI model, a product of intense research and dedication, is doing its best to sieve out the mundane and present you with only the best pictures. It aims to provide quality over quantity: if a bird's visit is fleeting and results in blurry shots, our AI will try not to include them in the postcard.
A photo will generally earn the AI's nod of approval only if it encapsulates an in-focus view of a bird where at least one eye and the beak are discernible. This rigorous selection process is the reason why, at times, you may notice the "taking postcards" status on the app but end up without any new postcards from that particular session.
The magic doesn't end once the photos are taken and selected. They are then fed into our bird detector model which can spot up to 6 birds and distinguish between up to 4 different species in a single shot. Each detected bird is further passed through our species classifier, identifying the exact species of the bird.
It's essential to understand that no system is flawless. However, leveraging machine learning ensures our model will keep getting better.
Session Length and Camera Cooldown
Our Bird Buddy feeders have been designed to deliver beautiful bird visitor moments. Here's how the system functions in real life:
Proximity Sensor Activation: When a bird lands on the feeder, the proximity sensor is triggered, launching a postcard session. This session can last anywhere from 30 seconds to 2 minutes, based on the bird's interactions.
Session Duration: If no bird activity is detected within the first 45 seconds, the session ends without any upload. However, if birds are present within this timeframe, the device captures videos and generates potential postcards.
AI Analysis: Post-session, our AI evaluates all the captures. It selects and uploads only those images that meet the quality standards. This means that even if multiple images are captured, only the best shots make it to the final cut.
Cooldown Phase: After each session, the camera enters a 10-minute cooldown mode. This period is essential for preventing camera overheating and conserving battery life. During this phase, no postcards will be generated, regardless of new bird activity on the feeder during this time. Bird Buddy users have some influence over this process with power modes.
Given this process, the bird you may noticed on the feeder that didn't get captured, might have visited during the cooldown period after a previous session was completed, or the images captured didn't pass the AI's quality threshold. We continuously work on firmware updates to enhance the sensitivity and quality of captures, striving to miss fewer visitor moments.
To ensure that our model has the finest bird photos from your Bird Buddy to choose from, consider these guidelines:
Fill the Seed Tray to the Brim: Max out the seed tray. When there's ample seed, birds are likely to land in ideal positions for the motion sensor and camera.
- Adequate Lighting and Orientation: Poor lighting can compromise photo quality. If Bird Buddy is in a low-light spot, consider relocating it. Need tips to optimize image clarity?
→ Check out our tips for optimal photo quality 📸🐦
→ Check out our ultimate placement guide
Bird Acclimatization: Birds can be wary of new objects. At first, they might briefly visit or avoid the feeder. Given time, they'll become more familiar and linger, allowing for better photos.
- Stay Updated: We're regularly refining our camera and the bird-detection AI. Always keep your camera updated with the latest firmware to capture the best images.
Obstruction Issues: Birds might occasionally block the camera or, post-rain, mud may cloud the lens. Regularly clean the camera lens with a dry cloth to maintain clarity.
Set Your Location: Our AI relies on your set location for better bird identification. Make sure it's accurate to improve photo results.
Sensor Sensitivity with Updates: With the newest firmware, the motion sensor is finely tuned and might be highly sensitive. If you notice birds visiting but aren't getting postcards, check for potential disturbances. Items like fluttering leaves on a nearby tree branch or any constantly moving objects in the camera's view could interfere. Eliminating such distractions can enhance photo capture.
if you are not receiving postcards, make sure the live stream is functional. If the live feed is operational, your camera should be working.
If you're not receiving any photos even though your Bird Buddy has been receiving regular visitors for a few days in a row anyway, please contact our support team.