The Listening Nest Box

Results

The box went up in October 2026, so the honest summary is: the technology works, the birds have not moved in yet.

What works

What is not proven yet

Lessons so far

Next steps

  1. Let it run over the winter and collect pictures (they are labelled bird and nobird).
  2. Keep tuning the thresholds with real data (see Detection), try the "empty after" time (300 seconds is a likely candidate) and review false alarms.
  3. If the model misses too many birds, retrain it with my own pictures, labelled locally on a computer, without any cloud service.
  4. Watch the IR module's temperature in summer.
  5. Maybe: the frogs in the garden pond, and an ultrasound microphone for bats.

How to tell a bird from a surveillance drone

A field guide for the careful observer. Check at least three before you report anything.

If in doubt, write to the address in the footer.

Black-and-white infrared picture from inside the nest box: a tit stands at the bottom edge and looks straight into the lens, with a dark eye and a white cheek; no detection frame
Exhibit A, 28 October, 08:53. The subject looks straight into the lens. The detection model scored it 0.00 and did not see a bird at all. Make of that what you will.

The first visitor

In October 2026 the first bird looked in. Only its head and beak reach into the picture at the top edge, which is exactly the situation that is hard for a general-purpose detection model, and the reason motion counts as well (see Detection).

Black-and-white infrared picture from inside the nest box: a small bird's head with a beak pointing down appears at the top edge
The first visitor (suspected drone): head and beak at the top edge of the picture.

More visits

On 17 October two more birds looked in, at 13:22 and 13:48. Both times only the head and beak reach into the picture at the top edge, and both times the model scored 0.00 while the motion check raised the alarm. This is exactly the case I want to tune in spring, for example by also looking at a strip along the top edge of the picture.

Picture from inside the empty nest box: a pale head with a dark pointed beak reaches into the picture from the top edge
A pale head and a pointed beak at the top edge. Model score: 0.00.
Picture from inside the empty nest box: a dark grey head with a small black beak reaches into the picture from the top edge
A dark head with a small black beak. Model score: 0.00.

False alarm: a patch of sunlight

Not every alarm is a bird. On the same afternoon the motion check fired again at 15:56, and the picture shows why: sunlight comes through the entrance hole and paints an orange patch on the wall, with a bright reflection at the top edge. The model found no bird (score 0.00), which is correct. The patch moves with the sun and changes the picture just like an animal would. In spring I will look at filters for this, for example ignoring bright patches that have no dark shape.

Picture from inside the empty nest box: an orange oval patch of sunlight on the right wall and a bright reflection at the top edge
A patch of sunlight on the wall. Motion alarm, model score: 0.00.