How many security cameras does a drugstore need?

Start from the drugstore model (a model that opens with a free account), or draw your own store. The entrance, the checkouts, the aisles between tall shelves: how many security cameras does a drugstore need, and where? And do ordinary cameras on the walls or all-round cameras on the ceiling see more? In this article, in a simulated drugstore, the AI asks suggest_cameras where to put the cameras and counts how much is seen for each number of cameras, mounted the two ways. It used the same tools an AI calls over MCP; every figure and picture below is their result.

1. What was checked

A simulated single-storey drugstore (36 × 24 m). On the sales floor, nine gondola shelves 1.8 m tall and 10 m long stand in north–south rows with the aisles between them; a fridge case runs along the west wall, three checkouts stand east of the entrance to the south, and the dispensary is to the east. To the north are the stock room, the office, the staff room and the toilets.

The AI gave suggest_cameras these conditions:

• What must be seen: the entrance with 2 m round it and the checkouts with 1.5 m round them count three times; the whole sales floor once.
• How they are mounted: two ways, compared (the camera figures are assumed): - Ordinary cameras on the walls: on the walls’ inner faces, 2.6 m above the floor; 90° across, 60° up and down, seeing 20 m. - All-round cameras on the ceiling: every 2 m on the ceiling (3.3 m above the floor), 360° and looking straight down, seeing 10 m.
• None in the stock room, the office, the staff room, the toilets or the dispensary.
• How it counts: for each point of a 0.25 m grid on the floor, whether it is in a camera’s view with nothing (shelves, walls) in between.

suggest_cameras adds the camera that sees the most first, then the next, then swaps one camera for another as long as that sees more. The result is not a proven optimum.

2. How much is seen, the two ways

• Ordinary cameras on the walls: 70.5% with 4 and 89.6% with 8. After choosing them, three swaps took them to 91.3% (by area, 90.7% of 496 m²).
• All-round cameras on the ceiling: 75% with 4 and 91.9% with 8. After choosing them, four swaps took them to 92.9% (by area, 91.8%).

The areas are counted on the 0.25 m grid, leaving out what is under the shelves and other parts.

• Both pass 90% with eight. The all-round ceiling cameras see more of the entrance and the checkouts.
• The eight on the walls: four on the inner face of the south (front) wall and one in the south-east corner, and three on the wall to the back rooms to the north. All look across the sales floor at an angle from the walls at its ends.
• The eight on the ceiling are spread over the sales floor.

get_view_image’s visibility draws what the eight placed wall cameras see. The figures in the strip above are for the whole floor, the back rooms included, not only what must be seen.

3. Do the toilets or the staff room show?

With the eight wall cameras placed, the AI measured the rooms to the north (with count: 0, suggest_cameras measures the cameras already placed):

• With the doors shut, no room shows. With them open, 0.4% of the toilets and 1% of the staff room show through the open doors.
• With the eight all-round ceiling cameras, 1.3% of the toilets shows with the doors open: the suggestion’s summary reports how much is seen of any room that looks like a toilet or a changing room.

4. How to use it, and what it does not tell you

• Compare the ways of mounting. For the same places to be seen, you can compare wall and ceiling mounting, and different kinds of camera. How far a camera sees depends on the model and the picture quality: give the figures of the cameras you have in mind.
• Measure the rooms that must not be seen too. Once the cameras are placed, count: 0 against the toilets and the staff room shows whether any camera sees into them, with the doors open and shut.
• Put up a notice. The Q&A of Japan’s Personal Information Protection Commission (Q1-13) says that a business that uses a security camera to take images in which a person can be identified should specify the purpose as far as it can and use the images only for it. Where people cannot easily tell that a camera is taking their picture, the business should 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 you:

• Whether a point is seen is judged from the field of view and the line of sight only. Image quality, light, and whether what hands are doing can be made out are not calculated.
• Customers and staff moving and hiding things are not counted.
• Shelves count as boxes that hide what is behind them up to their height; seeing between the shelves’ loads is not counted.
• The cameras’ fields of view and range are assumed; the all-round cameras’ “10 m” stands for how far their picture is good enough.
• The store is illustrative: the sizes are examples.

Tools used

• suggest_cameras: from what must be seen and its weights, where cameras may go (walls, ceiling), at what height and with which lens, it gives the places and the share seen for each number of cameras; with count: 0, how much the placed cameras see of a room
• edit_space: places the suggested cameras as camera parts
• get_visibility: checks what the placed cameras see
• get_view_image (visibility): draws what they see and the blind spots on the plan