Open the 3PL warehouse model (no account needed). Goods in and out at the docks, the receiving and dispatch areas, the rack aisles: how many security cameras does a warehouse need, and where? In this article, on the ground floor of a simulated 3PL warehouse, the AI asks the tool to suggest camera positions and counts how much is seen for each number of cameras and each mounting height. It used the same tools an AI calls over MCP; every figure and picture below is their result.
1. What was checked
The ground floor (72 × 48 m) of a simulated 3PL warehouse: six docks in the north wall (two inbound, four outbound) with the receiving and dispatch areas in front of them, the main aisle in the middle, pallet racks 7.5 m and 3.6 m tall in north–south rows to the west, packing to the east and returns to the south-east. In the north-west corner are the break room, the locker room, the office and the toilets.
The AI gave suggest_cameras these conditions:
- What must be seen: the six docks with two metres around them count three times; receiving, dispatch, packing and returns twice; the main aisles and the rack area once.
- Where cameras may go: on the walls and columns, 6 m above the floor; none in the north-west corner with the break room, locker room, office and toilets. That gives 202 candidate positions, with directions every 15° and downward tilts of 15°, 30°, 45° and 60°.
- The camera: 90° across, 60° up and down, with a range of 30 m (assumed).
- How it counts: for each point of a 0.5 m grid on the floor, whether it is in a camera’s view with nothing (racks, walls or goods) in between.
suggest_cameras adds the camera that sees the most first, then the next, then swaps cameras for others as long as a swap sees more. The result is not a proven optimum.
2. How much is seen, by the number of cameras
- Added one at a time, the cameras see 53.6% with 4 cameras, 76.8% with 8, 83% with 10, 87.7% with 12 and 92.9% with 16. Each camera adds less than the one before.
- After the sixteen were chosen, ten swaps took them to 94.3% (by area, 94.5% of 1,575 m²).
3. Where the sixteen go
| What must be seen | Area | Share seen |
|---|---|---|
| The docks and two metres around | 71.5 m² | 94.8% |
| Receiving | 79.5 m² | 85.2% |
| Dispatch | 159.5 m² | 98.4% |
| Packing | 37.5 m² | 93.3% |
| Returns | 82 m² | 89.6% |
| The main aisles | 641.3 m² | 97.3% |
| The rack area | 503.8 m² | 92% |
- All sixteen go on the outer walls; no column was chosen. The eight on the south wall look north, six of them along the rack aisles. The four on the north wall look at the receiving and dispatch areas from the dock side, and two each on the west and east walls look in from the sides.
- 3.1% of receiving, 2.7% of packing and 0.3% of the rack area are seen from no candidate position (the tool’s result lists them): these places stay unseen however many cameras are added.
get_view_image draws what the sixteen placed cameras see. The figures in the strip above are for the whole ground floor, the office and the toilets included, not only what must be seen.
4. Mounted higher
The AI asked again for ten cameras at 4 m, 6 m and 8 m:
| Height | Share seen (10 cameras) | The rack area |
|---|---|---|
| 4 m | 83% | 72.6% |
| 6 m | 83.4% | 73.5% |
| 8 m | 85.8% | 73.3% |
- The height changes the whole share by only 2.8 percentage points.
- The rack area stays in the 70s at every height. The racks are 7.5 m tall and the next aisle cannot be seen over them, so each aisle needs a camera looking along it.
5. Do the office, the toilets or the locker room show?
With the sixteen placed, the AI measured the rooms in the north-west corner (asked for zero new cameras, suggest_cameras measures the cameras already placed):
- The toilets, the locker room, the break room and the office are 0% seen, even with the doors open.
- The corridor in front of the office is 4.6% seen with the door to the warehouse floor open, and 0% with it shut.
6. How to use it, and what it does not tell
- Decide by the weights. Raising the weight of what must be seen most (the docks, receiving and dispatch) gives positions that favour them.
- Rack aisles, one by one. Height does not make up for tall racks: count a camera looking along each aisle from its end.
- Measure the rooms that must not be seen too. After placing the cameras, measuring the toilets and the locker room with zero new cameras shows whether any camera sees into them.
- Put up the notice. The Personal Information Protection Commission’s Q&A (Q1-13) says that a business taking images in which a person can be identified with a security camera should specify the purpose as far as it can and use the images within it, and, where people cannot easily tell that a camera is taking their picture, take steps such as a “security camera in operation” notice (Act on the Protection of Personal Information, Article 20(1)).
What it does not tell:
- Whether a point is seen is counted from the field of view and the line of sight only. Image quality, light, and whether labels and slips can be read are not calculated.
- Goods and forklifts moving and hiding things, and seeing through gaps between the racks’ loads, are not counted (racks count as boxes that hide what is behind them up to their height).
- The camera’s field of view and range are assumed: check them against the models you have in mind.
- The warehouse is illustrative: the sizes are examples.
Tools used
suggest_cameras: from what must be seen and its weights, where cameras may go and at what height, and the lens, it gives the positions and the share seen for each number of cameras; asked for zero new cameras, it measures how much of a room the placed cameras seeedit_space: places the sixteen suggested cameras as camera partsget_visibility: checks what the sixteen placed cameras seeget_view_image: draws what they see and the blind spots on the plan
