VisionConcept
ParkSense AI
A palm-sized edge box that plugs into your existing parking cameras.
- Counts free bays in real time
- Flags overstays and wrong parking
- Guides drivers to open spaces
From $499 per edge box
DetailsReservation not completed. Checkout was cancelled and you have not been charged. You can try again whenever you are ready.
IEMI Edge Hub + VisionCam
Eight cameras, one quiet box, no video leaves the building
Concept · not for sale
Engineering design: IEMI Edge Hub + VisionCamTarget price from$1,190LotWatch installed kit (proposal)
Software: $39 per site per month (proposal)
Concept, not for sale yet. Tell us if you want it; that decides what we build.
Rate this idea and tell us what to change
Drag to rotate · right-drag to move · scroll to zoom · click a part for its size and placeSwipe to rotate · pinch to zoom · two fingers to move · tap a part for its size
How it fits together
IEMI Edge Hub + VisionCam, part by part
Scroll or tap a part. Press “See the 3D CAD design” to watch the model, built from our CAD, light each part up and show how it connects.
Real-time 3D render · drag to rotate · final design may varySwipe sideways to rotate
01
The whole IEMI Edge Hub + VisionCam
Cameras send video to one Pi 5 hub; a Hailo chip finds cars, trucks and people on site, and only small events leave the building.
02Brain
Raspberry Pi 5 (8 GB)
The brain: it runs the edgehub vision + lotwatch program
03
Bought PoE camera
Outdoor camera, used as it comes, on its own camera network
Camera network · PoE Ethernet · RTSP · H.264 video substream
04
VisionCam-W
Cheap indoor camera for dock doors and shop floors
XIAO ESP32S3 Sense · Wi-Fi to hub access point · MJPEG video
05
VisionCam-Z
Small indoor H.264 camera with a wide lens
Pi Zero 2 W · CSI-2 ribbon, then RTSP · H.264 video
06
AI HAT+ (Hailo-8)
Runs the vehicle and person detector on every camera
Raspberry Pi 5 (8 GB) · PCIe Gen 3 (x1) ribbon · Frames in, boxes out
07
USB SSD
Keeps 7 days of event clips on site
Raspberry Pi 5 (8 GB) · USB 3 cable · Event clips
08
Coolers
Keep the Pi and Hailo chip cool inside the closed box
Raspberry Pi 5 (8 GB) · Fan header, GPIO 18 PWM · Fan power and speed
09
LotWatch LED sign
Shows free spaces on a street-side panel
ESP32-S3 (sign) · Wi-Fi · local MQTT · Free-space count text
10
All connected
8 key parts, one program, and your phone
What it is
An on-site AI box for up to 8 cameras that sends events, not video, starting with parking.
What it does
- Counts free parking spaces live
- Texts staff about overstays
- Keeps video on site
- Runs several IEMI app packs
Who it's for
Managers of 20–150 space lots, small warehouses and MachineState shops.
The idea
One quiet hub counts a small lot's free spaces from the cameras on site, lights the entrance sign and flags overstays
The problem today
Small apartment, clinic and hotel lots have no space count: drivers circle, visitor spaces fill with residents and the manager answers parking complaints by walking the lot. Parking guidance systems are priced for big operators, with a sensor per space and cloud video.
With IEMI Edge Hub + VisionCam
Two to four cameras the site already has, or cheap PoE ones, feed one hub indoors. It turns each space into free or taken, lights the count on a sign at the entrance and on a phone page, flags a visitor space held past its limit, and sends a weekly report. No video leaves the building.
- 1Edge Hub in the building's comms closetOne quiet box takes every camera's low-res stream, turns each space into free or taken and sends events, never video
- 2PoE camera on the building's east cornerLooks down the resident row and the island from 5 m up, so 15 to 40 spaces are read from one view
- 3PoE camera on the west light poleCovers the visitor row and the entrance lane; a space held past 2 h plus grace raises an overstay with one blurred snapshot
- 4Wi-Fi camera under the carport roofA small build camera where there is power but no cable run, reading the covered spaces the outdoor cameras cannot see
- 5Free-space sign at the entranceShows Visitor and Resident free counts within 30 s of a change, and FULL with the overflow street when the lot fills
How the parts connect
IEMI Edge Hub + VisionCam, wired up
Every part, what links it to the brain, and how the data reaches your phone. A solid line is a cable; a dashed line is a radio link.
The program
What controls IEMI Edge Hub + VisionCam
- Runs on
- Raspberry Pi 5
- Written in
- Python (GStreamer + HailoRT) and Node.js
- System
- Raspberry Pi OS Lite (64-bit), Docker
- AI model
- YOLOv8n compiled for the Hailo-8
- 1Read each camera's H.264 substream and drop to 2 fps
- 2Send frames round-robin to the Hailo-8 for detection
- 3Publish boxes on the local MQTT broker
- 4LotWatch maps boxes to space polygons with hysteresis
- 5The agent sends space changes and overstays to the cloud
pipe = Gst.parse_launch(
'rtspsrc location=rtsp://10.42.0.11/sub ! rtph264depay ! h264parse'
' ! avdec_h264 ! videorate ! video/x-raw,framerate=2/1 ! videoconvert'
' ! videoscale add-borders=true ! video/x-raw,width=640,height=640'
' ! hailonet hef-path=yolov8n.hef ! hailofilter ! appsink name=out')
pipe.set_state(Gst.State.PLAYING)
while True:
sample = out.pull_sample()
boxes = hailo_boxes(sample) # {cls, conf, x, y, w, h}
mqtt.publish(f'local/det/{cam}', json.dumps(boxes))
# lotwatch (Node) maps boxes to spaces and raises overstay eventsTo your phone
How the data reaches LotWatch
Manager's phone or laptop browser
- Hub
- Ethernet · MQTT/TLS (events only)
- IEMI cloud
- HTTPS web app, SMS
- Phone
What you see in the app
One screen for each job IEMI Edge Hub + VisionCam does.
Free spacesA visitor pulls into the apartment lot. The sign at the entrance and the phone page show how many visitor and resident spaces are free.
OverstayA car sits in 2-hour visitor space V07 past its limit. The front desk gets the space, the time and a snapshot with the plate blurred.
Weekly reportEvery Monday the manager gets one email with the hours the lot was full, the overstays and one suggestion for the spaces.
App screens with sample data. The interface is in development.
What the app gets
| Space state | V07 occupied since 13:41 | On change |
|---|---|---|
| Free count | Visitor 3 of 12 free | Every 60 s |
| Overstay alert | V07: 131 min, limit 120 | When limit passes |
| Camera health | OK, 10 fps in | Every 60 s |
| Hub health | SoC 61 °C | Every 60 s |
Video never leaves the site. Only events go out, and plate reading is off by default.
CAD design · Design complete on paper (stages 1–6, pilot rev A); nothing built, flashed or measured
Designed in code, checked in 3D
- 14 printed parts, each a single closed solid
- Every bought part fitted inside and checked for collisions
- Assembled exactly as shown in the 3D CAD view (the button on the picture at the top)
See the CAD parts render
Key specs
| AI processor | Raspberry Pi 5 8GB + Raspberry Pi AI HAT+ 26 TOPS (Hailo-8) |
|---|---|
| Camera | Bought Reolink RLC-510A 5 MP ONVIF PoE cameras outdoors; VisionCam-W (Seeed XIAO ESP32S3 Sense, wide lens) for indoor or cable-less spots |
| Sensors | Up to 8 camera streams · IEMI sensor nodes over the hub's Wi-Fi access point |
| Connectivity | Ethernet uplink, or optional USB LTE dongle · Isolated camera network on a TP-Link TL-SG1210P 8-port PoE+ switch · Wi-Fi access point for VisionCams · MQTT over TLS, outbound only |
| Power | Raspberry Pi 27 W USB-C supply; about 12 W typical (est.). Optional UPS HAT with 18650 cells. PoE switch on its own outlet |
| Enclosure | Indoor printed PETG hub box with a 40 mm lid fan, DIN-rail clip option; VisionCam in ASA with an outdoor hood and pole bracket |
| Dimensions | Hub 129 × 137 × 64 mm; VisionCam 80 × 48 × 30 mm |
| Ingress rating | None certified; VisionCam IP65 is design intent only |
All target specifications
Hub (design targets)
| Streams | 8 × 10 fps YOLOv8n (to be confirmed on the bench) |
|---|---|
| Backhaul | Under 100 MB a month |
| Video | 30 s ring in RAM only; event clips on the SSD |
| Offline | Events queue for 24 h or more |
LotWatch (targets)
| Lot size | 20–150 spaces with 2–4 cameras |
|---|---|
| Space coverage | About 15–40 spaces per camera |
| Accuracy | 97% per space over 1,000 day and 1,000 night samples |
| False overstays | At most 1 per 100 space-days |
Privacy
| Snapshots | One crop per overstay, plates blurred by default |
|---|---|
| Face recognition | None |
| Camera audio | Always off |
Targets for production units; they may change during development.
Pricing
Hardware: $1,190 LotWatch installed kit (proposal)
Software: $39 per site per month (proposal)
Proposal only; a demand test is needed. Parts cost about $411 for the hub core, $496 with the PoE switch, and about $745 for a typical LotWatch site with 4 cameras and a sign (est.). A VisionCam-W costs about $30 in parts.
Indicative launch pricing in USD, excluding taxes and shipping. Final pricing is confirmed before any balance is charged.
Community
What do you think of IEMI Edge Hub + VisionCam?
IEMI Edge Hub + VisionCam is in development. Rate the idea, tell us what to change about the features, specs or price, and we'll build it with your feedback.
How much people want it
–
No ratings yet
- 50
- 40
- 30
- 20
- 10
Follow IEMI Edge Hub + VisionCam
Get build updates and first access when pre-orders open.
Comments
No comments yet. Tell us what you'd want from IEMI Edge Hub + VisionCam.
Questions
Ask about IEMI Edge Hub + VisionCam
Tell us what this would save you, or ask anything. We reply within two business days.
Thanks, we have it.
A member of the IEMI Labs team will reply within two business days.
From the product lab · ideation phase
Lab updates for IEMI Edge Hub + VisionCam
LatestA 20 to 150 space apartment, clinic or hotel lot needs a live free-space count, overstay evidence and a weekly report; per-space sensor and cloud video systems are priced for big operators (sources in DESIGN.md).
- Drawn idea plan (idea.json): a 40-space apartment lot from above with the building, resident and visitor rows, a covered carport and the entrance; pins for the hub in the comms closet, two PoE cameras, a Wi-Fi camera under the carport and the free-space sign at the entrance
- The product picture is now this drawn plan instead of the plain fallback card; drawn from the rev A design notes, not the CAD
- What we kept from the market work: buy weatherproof PoE cameras for outdoors and build our own small camera only where there is no cable run
- Next: drawn plans for more catalogue products that still show the fallback card
IEMI Edge Hub + VisionCam FAQ
General FAQIs it available?
No. The Edge Hub and LotWatch are a concept in the in-house design and testing stage. The design is complete on paper and in CAD, but nothing has been built, flashed or measured. We are not taking orders; you can ask to join the pilot list.
Can I use my existing cameras?
Often, yes, if they offer ONVIF or RTSP. For new outdoor cameras we recommend a bought PoE camera such as the Reolink RLC-510A. We build our own VisionCam only for indoor or cable-less spots.
Does it work at night?
It needs IR or lighting, and night accuracy is our biggest open risk. If it does not reach the 97% target at night, spaces will show unknown rather than a guess.
Does it read licence plates?
Not by default. Plate reading is optional, reads only overstayed spaces, needs signage, and will not be offered to any customer before a lawyer has reviewed it.
Is it a security camera recorder?
No. It does not record continuously and is not a safety device. It keeps a 30-second ring in memory and saves only short event clips.
Related products
VisionConcept
WasteSort Coach
A clip-on camera coach for the recycling bins you already have.
- Lights the right bin for each item
- Says "not sure" instead of guessing
- Estimates each station's contamination
From $299 per 3-bin station
Details
VisionConcept
ForkGuard AI
A rugged camera for the forklift roof with an in-cab display that warns of pedestrians.
- Detects people near the forklift
- Beeps and flashes in the cab
- Logs anonymous near misses
From $799 per forklift
DetailsIEMI Edge Hub + VisionCam
Concept · target $1,190