Camera Project Acceptance Testing: Scenarios, Criteria and the Acceptance Record

Camera acceptance testing: image/DORI checks from recordings, retention and NTP verification, alarm, network and access tests, a test–criterion–evidence table and the structure of the acceptance record.
Technical summary
Technology
IP camera, NVR/VMS, NTP, RTSP, ffprobe, nmap, DORI, IEC 62676-4
Topic
Post-installation technical acceptance of a camera project, test scenarios and the acceptance record
Level
Intermediate
Prerequisites
  • Approved design and specification
  • Camera list (location, purpose, image target)
  • VMS/NVR admin access
In this article8 sections

When the installation team leaves the site, the system is assumed to work. Whether the night image is good enough, whether recordings are kept for the planned period and whether they can be exported with the correct time during an incident, however, often only come to light at the first incident, that is, at the worst moment. The only tool that shows this in advance is a structured acceptance test. Below we define, for the client, contractor and consultant, the test scenarios, the criterion and evidence for each test, and the structure of the acceptance record.

Purpose and parties of the acceptance test

In enterprise camera systems, the purpose of the acceptance test is to show with evidence that the system meets the requirements written in the specification. Two consequences follow from this definition. If there are no measurable targets in the specification, no test can be done. An observation that “it works” isn’t evidence; a recorded screenshot, an exported video or a log is.

The parties and their roles are usually distributed like this:

  • Contractor: Prepares the test plan, runs the tests, collects the evidence and fixes nonconformities.
  • Client security unit: Verifies that image targets and operational scenarios are met.
  • Client IT unit: Verifies network, access, security hardening and monitoring integration.
  • Consultant or supervising engineer (if any): Oversees the test method and the impartiality of results.
  • Data protection officer: Checks that privacy masks, retention period and access permissions comply with policy.

Not inviting the IT unit to the test at all is a common gap on site; network and password items then pass without anyone’s signature. Analytics accuracy tests can be part of this general acceptance but have their own method; for measuring precision and recall, see the video analytics performance testing article.

Preparing the test plan

The test plan should be approved by the parties before installation is finished. A plan written after installation tests what was installed; it doesn’t compare it with what was requested. The backbone of the plan is the camera list prepared during design. For each camera there are at least these fields: camera ID, location, purpose, target image level (detection, observation, recognition, identification), target distance, recording mode and retention period. How to build the list is in the IP camera system design guide.

Other sections the plan should include:

  • Test timing: Day and night tests are done in separate sessions. The night test should be done in real darkness with site lighting in its normal operating state; a night test done with all floodlights on proves nothing.
  • Test equipment: Laser meter, test target (a person or resolution chart), laptop, PoE tester and, if needed, a lux meter.
  • Acceptance criteria: The “pass” condition for each test is written in advance; for example, the time synchronization tolerance or the retention period.
  • Nonconformity classes: Definitions of “major” nonconformities that block acceptance and “minor” ones that don’t block acceptance and are to be fixed within a set period.
  • Sampling rule: In projects with many cameras, some tests can be done by sampling; which tests are done on every camera and which by sampling is written down.

Image and coverage tests

The first rule of image testing: assessment is done from the recording, not on the live image. While the main stream is watched in live view, recording may be done from the substream or at a different bitrate. A face that looks sharp on the monitor may have broken into blocks in the recording. What will be used at the time of an incident is the recording.

  1. Coverage: The frame is compared with the coverage drawing in the design. That defined targets such as doors, docks and checkouts are in the frame is documented with a screenshot.
  2. Image level: A test person walks toward the camera at the target distance measured with a laser meter. Whether the target level is reached is assessed on a frame taken from the recording. Pixel density can also be estimated from the number of pixels a known width of the person covers in the image.
  3. Night performance: The same test is repeated at night. IR reflection, overly bright faces, blur on moving targets and noise are checked.
  4. Backlight: For cameras at glass façades or entrances facing the sun, WDR performance is tested at the critical time of day; a test done at noon doesn’t show the morning sun.
  5. Privacy masks: It’s verified that masks are in the right place in the live image, the recording and exported video.
  6. Stream parameters: Codec, resolution and frame rate are checked against the specification.

To verify stream parameters without going into each camera interface one by one, ffprobe from the FFmpeg package can be used. The RTSP path varies by manufacturer and stream profile; the address below is an example:

ffprobe -v error -rtsp_transport tcp \
  -show_entries stream=codec_name,width,height,avg_frame_rate \
  -of compact "rtsp://user:[email protected]:554/<stream-path>"

Recording, retention and time tests

Recording tests show that the system can produce evidence at the time of an incident. The most debated item is the retention period, because the existence of 30 days of recording can’t be physically proven right after installation.

  • Recording continuity: During the test period (for example a 72-hour trial run), the recording timelines of all cameras are examined for gaps.
  • Retention period: The expected retention period is calculated from the actual daily data volume measured during the trial run and the usable disk capacity. Final verification is done by looking at the date of the oldest recording at the end of the first month of operation, and this condition is written explicitly into the record. For the calculation method, see the how many days does an NVR record article.
  • Time synchronization: It’s verified that all cameras, NVR/VMS servers and viewing stations are connected to the same NTP source and that the difference stays within the tolerance in the specification. A single camera whose clock has drifted by a few minutes makes the whole timeline disputable in an incident review.
  • Export: An incident recording is exported; it’s shown that it opens with the player, that the timestamp is correct and that digital signature or watermark verification, if any, succeeds. The topic is covered in detail in the video recording integrity and digital signatures article.
  • Failure behavior: It’s tested how recording continues, and whether an alarm is raised, when a disk is removed or, in a redundant setup, a recording server is shut down. If the disk failure alarm reaches no one, the redundancy exists only on paper.

Example commands, by operating system, to verify time synchronization on the VMS server:

# Windows Server (W32Time)
w32tm /query /status
w32tm /stripchart /computer:10.20.0.10 /samples:5 /dataonly

# Linux (chrony)
chronyc tracking
chronyc sources -v

Network, access and security tests

Network and security tests are the IT unit’s responsibility and the most often skipped part. Minimum scope:

  • Segmentation: It’s tested that cameras are on the defined VLAN and that direct access to cameras from the user network is blocked. If the camera interface opens from a laptop on the user network, there’s no segmentation.
  • Open services: The camera network is checked with an authorized scan; it’s verified that unused services such as Telnet, UPnP and P2P cloud services are off. The scan is done only with the client’s approval and within a defined time window.
  • Accounts: It’s checked that default passwords have been changed, personal user accounts and roles have been defined, and no shared account is used.
  • Firmware: That all devices are on the current, approved firmware version as of the delivery date is documented with the inventory.
  • Remote access: It’s verified that remote access is possible only by the approved method (for example VPN) and that no port forwarding has been set up.
  • Power: A controlled test shows that the UPS powers the recording system and PoE switches for the expected time during a mains outage.
# Service discovery on the camera VLAN (authorized scan)
nmap -sV -p 21,22,23,80,443,554,8000,8080 10.20.30.0/24

# UDP: SSDP/UPnP (1900) and WS-Discovery (3702)
sudo nmap -sU -p 1900,3702 10.20.30.0/24

For detailed hardening items, the IP camera and NVR cybersecurity checklist can be used alongside the acceptance test.

Test–criterion–evidence table

The table attached to the record defines the criterion and the form of evidence for each test. Adapt the example below to your specification:

Test Criterion Evidence
Coverage Consistent with the framing in the design Screenshot per camera
Image level (day/night) Target DORI level at the target distance Frame taken from the recording, with a distance note
Recording continuity No gaps during the trial run Timeline screenshot
Retention period Number of days in the specification Calculation sheet, end-of-first-month check record
Time synchronization The tolerance in the specification Command output, camera clock screens
Export Opens; signature verifies Exported file, verification screen
Alarm Disk, video loss, tamper notifications Notification record, email or event log
Network and security Services closed, passwords changed Scan report, account list
Power The defined duration on UPS Test time log

Acceptance record and handover documents

The acceptance record brings the test results and the conditions of acceptance together in a single document. The structure we recommend:

  1. Scope: Project name, locations in scope, device counts, specification and test plan references.
  2. Participants: The parties who took part in the tests and their roles.
  3. Results table: A pass, fail or conditional result for each test heading.
  4. Nonconformity list: For each nonconformity, the description, class (major/minor), owner and resolution date.
  5. Acceptance decision: Acceptance, conditional acceptance (with minors to be resolved) or rejection. Items to be verified later, such as the retention period, are stated explicitly here.
  6. Signatures and attachments: The test–criterion–evidence table and the location of the evidence archive.

Handover documents are part of acceptance. Missing documents mean missing work:

  • As-built design: camera locations, view directions, cable routes.
  • Device inventory: model, serial number, MAC and IP address, VLAN, firmware version, switch port.
  • Camera, NVR/VMS and switch configuration backups.
  • Handover of admin passwords by a secure method (password vault or sealed envelope); not sent by email.
  • User and role matrix, retention period settings.
  • Operator and administrator training records.
  • Warranty documents, maintenance plan and support contact details.

Acceptance of the cabling infrastructure is a separate job and must be finished before camera acceptance; for measurement and reporting principles, see the structured cabling specification and handover acceptance guide. For preparing the test plan and running independent acceptance, see our camera project management and acceptance testing page; for writing measurable targets into the specification, our IP camera system design page.

Frequently asked questions

Can a camera acceptance test be done on the live image?

The image quality decision is made from the recording. Live viewing may use a different stream or bitrate; what you’ll have at the time of an incident is the recording. The live image is used only as an aid for checking framing and coverage.

How is a 30-day retention period verified after installation?

The expected period is calculated from the daily data volume measured during the trial run and the usable disk capacity. Final verification is done by looking at the date of the oldest recording at the end of the first month of operation; write this condition explicitly into the acceptance record.

Which nonconformities block acceptance in the acceptance test?

Which nonconformities count as major is defined in advance in the test plan. Loss of recordings, failure to reach the target image level, time synchronization errors and unchanged default passwords are typically items that block acceptance.

Which documents should be attached to the acceptance record?

The test–criterion–evidence table, the nonconformity list, the as-built design, the device inventory, configuration backups, the user and role matrix, training records, and warranty and maintenance documents.

Support

Contact us for support or a discovery session on this topic.

Contact form
[email protected]

Eretek Video Surveillance

IP cameras, recording and central management infrastructure. We plan blind spots, retention periods and the camera network around the risks of each site.

Knowledge Center

Other articles