# Sampling or inspecting every egg: what does each one guarantee? > Full inspection checks every egg that passes the camera; sampling checks a chosen group meant to represent the batch. The two complement each other. For electronic candling equipment, MAPA's FAQ on Ordinance 612 requires a manual check by validated sampling; the same logic is good practice for a surface camera, with a record of every check. Updated: 2026-10-09 · URL: https://vigiovo.com/en/resources/full-inspection-vs-sampling · Português: https://vigiovo.com/recursos/inspecao-100-por-cento-vs-amostragem.md ## Key takeaways - Full inspection decides every egg; sampling checks the result and verifies the camera itself. - For electronic candling equipment, MAPA's FAQ requires a manual check by validated sampling (question 34); a surface camera isn't candling, but benefits from the same verification. - The camera applies the same criteria to every egg; the sample brings human checking and candling to the verification. - Use both together: the camera grades, and sampling confirms the camera keeps getting it right on your line. ## What is full inspection and what is sampling? In full inspection, every egg that passes the camera receives a class and a record. No egg goes undecided, except the unverified ones, which go to manual review. In sampling, the team sets aside a group of eggs and inspects that group in detail, with candling or visual inspection. The result applies to the group and does not guarantee that a specific egg is correct. That is why the sample has to be chosen with a method and recorded the same way every time. ## Full inspection or sampling: which covers more? The table compares the two approaches. The camera column describes [Vigiovo](https://vigiovo.com/en/how-it-works); the sampling column describes the manual procedure in general. The reading is practical, except the regulatory row, which follows the FAQ cited. **Camera inspection of every egg compared with sampling and manual checks** | Criterion | Full camera inspection | Sampling with manual checks | | --- | --- | --- | | Coverage | Every egg that passes the camera | A group of eggs chosen to represent the batch | | What it sees | The surface: dirt, visible cracks and shape | Includes the interior if the check is done by candling | | Hairline cracks | Not seen; needs candling or acoustic testing | Seen only if the sampled eggs are checked by candling or acoustic testing | | Consistency | Same result for the same image and configuration | Depends on who checks and how the sample was chosen | | Record | One record per egg, with class and reason | A record of each check, with frequency and method defined | | Role in the rules | A surface camera is not electronic candling; question 34's rule serves as a verification model | For electronic candling, the required manual check (question 34); for a camera, good practice | ## Why does sampling stay necessary alongside the camera? The MAPA FAQ on Ordinance 612 covers electronic candling in question 34. It says that electronic equipment giving results equivalent to traditional candling may be used, provided a manual check by a validated sampling system is planned (art. 26, § 4º). The same answer asks the company to describe, in its self-control programme, how it verifies the equipment: the frequency, the method of evaluation and the way the check is recorded. In other words, sampling needs a written rule, and cannot depend on someone remembering to do it. Sampling also answers a question the camera cannot answer alone: whether it is still getting it right on today's line. A deterministic camera gives the same result for the same image, but that proves consistency, not that it is calibrated for your shells. Calibration with images from your own conveyor is what ties the result to your line. ## How do you combine full inspection and sampling on one line? The combination works like this: the camera grades every egg and records each decision, and the team checks a sample on a planned schedule. 1. Set the frequency of the manual check and write it into the self-control programme. 2. Choose the sample so it covers different shifts, houses and batches, not only the start of the morning. 3. Check each sampled egg with the agreed method and compare the result with the camera's class. 4. Record each disagreement with its reason, and use the record to adjust the calibration or the checking. With the record of disagreements, the team can tell whether the error is in the camera, the lighting or the manual check. Without that record, every difference turns into a matter of opinion. ## What are the limits of each approach? Full inspection does not see what lies beyond the visible shell, so hairline cracks and internal defects are outside its reach, and it depends on calibration on your line. The accuracy figures published come from synthetic scenes with ground truth. Your line's accuracy is only known after calibration with real images. Sampling, for its part, only says something about the batch if the sample is well chosen and the check is always done the same way. A small sample always taken at the same time of day tells you little about the rest of the day. ## Frequently asked questions ### Can a camera replace sampling? The MAPA FAQ does not treat the camera as a replacement for sampling. It accepts electronic candling equipment provided a manual check by validated sampling is in place, described in the self-control programme (MAPA, question 34). In practice, the camera and the sampling work together. ### How many eggs should go into the sample? The FAQ cited does not set that number. The size and frequency of the sample must be defined and validated by the company itself, and described in its self-control programme together with the method of evaluation and the way it is recorded. ### How should I start combining the two approaches? Start with per-egg records and a planned manual check, during a pilot on your own line. Compare the two over a few weeks, note the disagreements, and only then adjust how often the manual check happens. ## Sources - [Perguntas e respostas sobre a Portaria MAPA nº 612/2022 (versão de 05/06/2023)](https://wikisda.agricultura.gov.br/manual_ovos/perguntas_e_respostas_port612_2023.pdf), Ministério da Agricultura e Pecuária (MAPA) - [Como o Vigiovo classifica ovos sem IA / How Vigiovo grades eggs without AI](https://vigiovo.com/como-funciona), Vigiovo ## Read next - [Egg grading line audit checklist](https://vigiovo.com/en/resources/egg-grading-line-audit-checklist.md) (Checklist, 2026-10-09): A checklist for auditing an egg grading line, with every item tied to the RIISPOA article or MAPA answer that supports it. - [Camera inspection pilot readiness checklist](https://vigiovo.com/en/resources/camera-pilot-readiness-checklist.md) (Checklist, 2026-10-09): Checklist for preparing an egg inspection pilot with a camera: camera, light, labelled eggs, computer, manual review and success criteria. - [Dirty eggs: what counts as dirty and how to grade them](https://vigiovo.com/en/resources/dirty-egg-grading.md) (Guide, 2026-10-09): When an egg counts as dirty, whether it can be washed, where it goes under Brazilian and EU rules, and how to measure shell dirt with a camera. ## How Vigiovo helps Vigiovo grades eggs on the conveyor as clean, dirty, cracked or unverified, in real time, with deterministic computer vision: every decision comes with the measurements behind it, on an ordinary CPU. https://vigiovo.com/en/how-it-works.md 4 min read.