Vehicle cameras that find roadside dumping spots and prove they stay clean
TL;DR Wardscan puts cameras on vehicles that already cover every street, such as collection trucks and sweeping vans, and uses image models to find roadside dumping spots, then logs when each one is cleared and whether it comes back. City corporations pay per ward. It answers a scored Swachh Survekshan indicator.
Indian cities are judged on how clean they look. The Swachh Survekshan 2024 toolkit gives 20 marks for having no garbage vulnerable points, the corners where people keep dumping waste, and another 20 for having no large street bins at all[1].
Cities fight these spots with drives and paint. Hyderabad counted 1,116 of them in 2017 and declared all cleared using signboards, wall art and rangoli[2]. Gurugram ran a ten-day drive in October 2026 to clear 21[3]. But drives are periodic. Nobody checks each spot every day, so a cleared corner can quietly fill up again and the city only finds out at the next inspection or complaint.
Wardscan fits a small camera and edge unit on vehicles that already pass every street daily. An image model flags piles of mixed waste on the roadside, tags the location and adds it to a ward map. Each spot gets a history - first seen, cleared, seen again. Ward officers get a morning list of new and returning spots with photos. Supervisors see which wards clear spots within a day and which do not. Faces and number plates are blurred on the device before anything is uploaded.
I reframed the seed, which pitched real-time AI monitoring of bins. Indian cities are scored for removing big street bins in favour of door-to-door collection[1], so a sensor in every bin works against the policy. The waste that escapes the system is on the roadside, and that is what Wardscan watches.
The buyer is the municipal corporation's health or sanitation department, paying a per-ward or per-vehicle subscription through a tender or GeM. Bengaluru's civic body has already earmarked ₹40 crore for AI analysis of camera feeds to detect garbage piles, potholes and similar problems[4], which shows that a budget line exists. Private collection contractors are a second buyer, since many contracts include penalties for uncleared spots.
Survekshan scoring keeps rewarding cities that clear dumping spots and remove bins[1]. The SWM Rules 2026 add a central online portal to track waste from collection to disposal[5], which pushes cities toward digital proof. US cities now buy this as a service: Dallas signed a camera-analytics startup in late 2025 to put two cameras on each of 50 trucks[6].
No public figure exists for spending by Indian cities on dumping-spot detection. There are two firm signals - Bengaluru's ₹40 crore allocation for AI camera analytics across several civic problems[4], and Dallas's contract worth about US$2.6 million (about ₹25 Cr) over three years for 50 trucks[6]. Both are adjacent, and neither is a market size.
That US startup sells vehicle-camera detection of illegal dumping, graffiti and overgrowth to cities[6]. A US waste-tech platform's cameras on trucks have flagged potholes, overflowing bins and dumping, with 954 potholes logged in six months in Kansas City[7]. In India, Bengaluru plans to use its Safe City CCTV network[4], which sees junctions but not back lanes. I found no Indian vendor selling ward-level dumping-spot tracking as a service.
| Difficulty | Hard (52/100) |
| Starting capital | ₹80 L–1.6 Cr |
| Time to launch | 9–12 months |
Detecting a waste pile on an Indian street is harder than it sounds, because construction debris, parked carts and festival decorations all confuse a model. The tougher part is selling to city corporations, where tenders take months and payments can lag a year. Ward officers also have to act on the list, or the data changes nothing at all.
No sector licence. Detecting and reporting dumping spots needs no permit, but procurement and data rules apply.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't trust the model on its own for penalties. Debris, carts and decorations confuse it, so a person should confirm any spot that leads to a fine.
Want to work out the AI layer for Wardscan? Talk to Yash →
Here's what can go south -
I think this suits a founder with computer-vision skills and patience for government sales, ideally with a contractor partner who already runs the vehicles. If you need revenue within six months then it's a poor fit. First prove that, in one ward, the model finds 85% of spots an inspector would mark and that officers clear most of them within 48 hours.
Plan for roughly ₹80 L–1.6 Cr for cameras, model training and the first city pilot.
No sector licence. You need GeM or tender registration to sell to cities and careful DPDP handling of street images.
Yes, US cities such as Dallas already buy this as a service. Check local privacy law on street cameras before you deploy.
The free Wardscan validation kit has the line-by-line capital breakdown, the effort scoring, the procurement and privacy checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for ward officers and contractors, and 5 assumptions to test. Download the kit →
Thinking about building Wardscan or something close to it? The make-or-break work is winning the first city pilot and getting ward staff to act on the daily list, and I can help you plan that first pilot. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.