Sewer CCTV Inspection Software: How to Turn Thousands of Hours of Footage Into a Rehabilitation Plan

Why sewer inspection data is nearly worthless in the format it arrives

A collection system manager takes delivery of a hard drive. On it are 640 inspections from a contractor's spring campaign: a WinCan database, a folder of MPEG files totalling most of a terabyte, a photo folder, and a spreadsheet the project engineer made because the database was missing upstream manhole IDs on 40 runs. Last year's contractor used ITpipes. The year before, a different crew delivered POSM files. The asset register lives in the GIS and uses a segment naming convention that none of the three contractors followed exactly. Somewhere in those inspections is the information that would tell the city which 3 percent of its pipe to rehabilitate first, and the practical way to get at it today is to have an engineer open videos one at a time.

The tools are not the problem in the way people assume. WinCan, ITpipes, POSM and CUES Granite XP are all competent inspection software and each does what it was built for, which is capturing and coding an inspection in a truck. Autodesk Info360 Asset sits on the analysis end and does real work there. The gap sits between them: nothing owns the pipeline from a contractor's raw deliverable, through validation against your register, through consistent scoring, to a rehabilitation decision you can put in front of a council or a regulator.

The stakes are set by the Clean Water Act. Sanitary sewer overflows attract enforcement, and enforcement produces consent decrees that specify inspection mileage, response times and rehabilitation commitments with deadlines and stipulated penalties attached. Once you are under a decree, your inspection data is not an engineering resource, it is evidence of compliance, and the format it arrives in becomes your problem rather than the contractor's.

Problem 1: contractor deliverables never match your asset register

The contractor inspects what they can access. Your register describes what you believe you own. These disagree constantly: a segment that was split when a manhole was added in 1996, a run inspected in reverse with the upstream and downstream nodes swapped, a lateral coded as a mainline, an ID that got typed with a leading zero this time and without one last time, a pipe that appears in the field and not in the GIS at all.

Inspection software validates within its own database, not against your register, because it does not have your register. So the reconciliation happens when someone loads the deliverable and starts finding mismatches, usually months after the crew has left the site and long past the point where a question can be answered by going back.

What a custom build does: validate at intake, before acceptance. The deliverable is parsed, every inspection is matched to a segment in your register by ID, by geometry, and by upstream and downstream node pairing, and anything that fails goes to an exception queue with the specific reason. Length discrepancies beyond a tolerance flag automatically, since a run inspected at 340 feet against a register length of 290 is either a bad record or a wrong pipe and both matter. The contractor gets the exception list while the crew is still in town. That single change, moving validation from months later to intake, is usually the highest value thing in the first release.

Problem 2: PACP coding is applied inconsistently and the score inherits every judgement

NASSCO's PACP gives the industry a shared vocabulary for defects and a structural and maintenance grading scheme built on it. It works, and certification means an operator has been trained. It does not mean two certified operators code the same pipe the same way. One codes a crack as CL, another as CM. One records every root intrusion, another records the significant ones. One records a joint offset at every joint on an old clay run, another summarises. The segment grades that come out the other end then get compared as if they were measurements.

The result is that a scoring system built on aggregate grades quietly ranks your pipes by which contractor inspected them. Nobody intends this and almost every system with multiple inspection vendors has it.

What a custom build does: keep the raw observations rather than only the rolled-up grades, then run quality checks across operators and contractors. Observation rate per hundred feet by pipe material and age, by operator, is a blunt instrument that finds systematic under-coding and over-coding quickly. Scoring is then computed by your own logic from raw observations, versioned, so when you refine the logic every segment rescores and you can see what moved. Automated defect detection from video is genuinely improving and worth using as a second opinion that flags disagreements with the human coding, not as a replacement for it. Used that way it finds the runs a tired operator skimmed at the end of a shift.

Problem 3: the video is terabytes and nobody can find the twelve seconds that matter

Every defect observation has a distance along the pipe and a timestamp in a video file. In practice the file is on a hard drive in a cabinet, or on a share nobody has mapped, and getting from a defect in a report to the picture of it takes fifteen minutes and a phone call. So engineers make rehabilitation decisions from the coded observations alone, and the video, which is the actual evidence, functions as an archive nobody opens.

What a custom build does: ingest video to object storage with a durable link from every observation to its exact timestamp, then serve it in the browser with adaptive streaming so a reviewer clicks a defect and sees it in two seconds from a laptop or a phone at a job site. Retention policy is worth designing deliberately: full video for a defined period, keyframes at every observation retained permanently. The keyframe decision alone changes storage cost by an order of magnitude and preserves the evidence you actually reuse. Once video is genuinely accessible, engineers start reviewing before committing to a rehabilitation method, and the number of surprises during construction drops.

Problem 4: a condition score is not a rehabilitation decision

A grade 5 structural defect in a 6 inch clay line behind a residential garage is a different problem from a grade 4 in a 36 inch interceptor under a hospital access road. Consequence of failure, groundwater conditions, whether the street is scheduled for paving next year, whether the segment is a known inflow contributor, and whether the pipe can be lined or must be dug all belong in the decision. Condition alone ranks pipes by how bad they look, which is not the same as which to fix.

Info360 Asset and similar platforms model risk and do it credibly. Where they tend to stop is your own local reasoning: this basin surcharges in wet weather so infiltration matters more here, this material and vintage combination has failed early for us repeatedly, this corridor is under a paving moratorium for five years after the last resurfacing so the window is now or 2031.

What a custom build does: a scoring model with condition, consequence and opportunity as separate inputs you can weight and change, producing a ranked list with the reasoning visible per segment. Rehabilitation method feasibility comes in as constraints, since a segment with severe deformation is not a lining candidate no matter how convenient that would be. Packages get assembled geographically so the ranked list turns into buildable projects rather than a list of scattered pipes. And every version of the ranking is retained, so when someone asks why a segment moved down the list between last year and this year, the answer is available.

Problem 5: a consent decree measures your programme, not your intentions

Under a decree you have inspection mileage commitments, defect response times, rehabilitation volumes, and reporting on a fixed schedule. The reporting is assembled by an engineer from several systems, checked twice because the consequences of an error are real, and it consumes weeks per cycle.

What a custom build does: track commitments as tracked objects with progress computed continuously from the inspection and construction records, so the position is known on any given day rather than at report time. Response clocks start automatically when a defect above a defined severity is accepted, and they escalate. The periodic report becomes a generated document with the underlying records one click away. Utilities under decrees describe this as the difference between reporting and knowing.

What this costs and how long it takes

Across the 2,000-plus projects Digital Heroes has delivered, here is the honest shape. A focused first release covering multi-format deliverable ingestion, validation against your asset register, raw PACP observation storage with coding quality checks, and a versioned defect score per segment runs $65,000 to $140,000 and ships in 12 to 18 weeks. A full platform adding browser video streaming with observation seek, rehabilitation decision modelling with method feasibility, geographic project packaging, contractor performance tracking and consent decree reporting runs $160,000 to $400,000 phased over 6 to 12 months.

What pushes cost up in this category: the number of distinct inspection software formats you have to accept, since each exchange format is real parsing work. Video volume and retention policy, which drives infrastructure cost directly and should be decided early rather than discovered. Manhole and lateral inspections alongside mainlines, which bring their own coding schemes. Integration with an existing work order platform such as Cityworks, and with a hydraulic model if you want inflow and infiltration analysis connected. And the condition of your asset register, which is the honest schedule risk: if a third of your segments cannot be matched confidently, that is register cleanup and it is real work.

What keeps cost down: begin with one contractor's format, mainline inspections only, and one basin. The register problems will surface immediately and cheaply.

Build versus buy, and when buying is right

Buy if you have one inspection crew using one software package on a system of modest size. WinCan or ITpipes already holds your data in a consistent format, your operators code consistently because there is one of them, and a custom build would be solving a problem you do not have.

Build when two or more of these are true. You contract inspection out to multiple vendors who deliver in different formats. Your system is large enough that manual reconciliation of deliverables consumes a staff position. You are under a consent decree or expect to be, which makes reporting and provable programme progress a first-order requirement. Your rehabilitation programme is a substantial recurring capital line and prioritisation decisions are currently argued rather than computed. Or you have years of historical inspections in incompatible formats that you cannot analyse together, which is the most common trigger we see.

Our position: inspection software is for the truck, and it is good at that. Condition assessment is for the utility, and it is a different system with different requirements. Trying to make truck software into an asset management programme is how utilities end up with four disconnected archives and an engineer opening videos one at a time.

How to choose a developer for sewer inspection software

Ask them how they will match a contractor's inspection to your asset register when the IDs do not agree. A serious answer covers geometry matching, node pairing, length tolerance, and an exception queue. An answer about importing data means they have not seen a real deliverable.

Ask whether they intend to store raw PACP observations or only the rolled-up grades. Only the grades means you can never change your scoring logic without re-inspecting, which is a decision that will haunt the programme for a decade.

Ask about video specifically: storage cost at your volume, retention policy options, and how fast a reviewer gets from a defect row to the frame. If they have not thought about keyframe extraction, they have not costed the platform honestly.

Ask what they have integrated. Cityworks, Esri, a hydraulic model and the various inspection exchange formats are four separate problems. Ask for the specific product and the specific format.

Ask who owns the code, the infrastructure and the video storage accounts before kickoff. At Digital Heroes the client owns the code from the first commit, which matters when the archive is a decade of legally significant evidence.

Start by handing a developer last season's deliverable exactly as the contractor sent it and asking them to tell you how many inspections cannot be matched to your register. That number is the project brief.