Two different quantities
Density is mass per unit volume — how much material is packed into a given space. It is reported in units like g/cm³ or kg/m³ (aluminum is about 2.7 g/cm³, steel about 7.8 g/cm³) and you measure it by weighing a known volume.
Porosityis the fraction of a material’s volume that is empty space — the voids, pores, and channels between the solid. It is reported as a percent or a decimal fraction. A dense metal sits near 0% porosity, structural foams reach 80–95%, and bone runs roughly 5–30% depending on type.
The two are linked but distinct. More void space means less mass in the same envelope, so higher porosity generally lowers bulk density. But porosity says nothing about what the solid is made of: two scaffolds with identical 60% porosity have very different densities if one is titanium and the other is PLA. You cannot infer one from the other without more information.
The relationship — and the one formula that connects them
When you know the true (solid) density of the material — the density of the fully dense solid with no pores, written ρs — porosity and bulk density are tied by a single relationship:
- ρbulk = ρs × (1 − porosity)
- equivalently, porosity = 1 − (ρbulk / ρs)
Worked example: a porous PLA scaffold has a true density of 1.25 g/cm³ and a measured bulk density of 0.50 g/cm³. Its porosity is 1 − (0.50 / 1.25) = 0.60, or 60%. Run the inverse — porosity to estimated bulk density — with the Porosity Calculator.
What image analysis actually measures
Image analysis does not weigh anything, so it cannot read density. What it measures is geometry: it separates pores from solid in a digital image and computes the void area fraction in the plane of that image. The workflow is three steps:
- Segment pores from material. Pores typically appear darker than the surrounding solid in a polished cross-section, so a grayscale threshold — or an AI segmentation model for noisy or low-contrast samples — labels each pixel as pore or solid.
- Measure pore area. Inside a defined region of interest, sum the area of every segmented pore (and, if you want a distribution, the area of each individual pore).
- Compute the fraction. Areal porosity = (pore area ÷ region area) × 100%. The same logic extends to 3D when you have a volumetric scan (micro-CT), where it becomes pore volume ÷ total volume.
This is exactly the pipeline behind ConductVision’s surface coverage & porosity and coating thickness & porosity workflows: draw an ROI, set the threshold, and export porosity percentage, pore count, and an overlay mask for review.
Does the 2D number equal the true 3D porosity?
This is the assumption every image-based porosity measurement rests on, and it is worth stating plainly. By the Delesse principle of quantitative stereology, the expected area fraction of a phase on a random plane through a material equals its volumefraction (Russ & DeHoff, 2000). That is what lets a flat micrograph stand in for a 3D void fraction — and it is the basis of standardized methods like manual point counting (ASTM E562) and area-percentage porosity for coatings (ASTM E2109).
The catch is the word expected. The equality holds on average, for a representative and unbiased section. A single field of view, an anisotropic pore network (aligned channels, layered porosity), or a sectioning artifact can push a single 2D number away from the true bulk porosity. The practical fix is built into the standards: measure multiple fields, report the mean and the spread, and treat one image as a sample, not the answer.
Calibration: pixels vs. real units
Porosity as a fraction is dimensionless — it is the same whether you count pixels or microns, because it is a ratio of areas. But the moment you want pore sizes (mean pore diameter, a pore-size distribution, D-values) in microns rather than pixels, you need a scale calibration: a scale bar, a feature of known size, or the imaging system’s pixel pitch. Without it, areas and diameters stay in pixels. Set the scale before reporting any dimensioned result.
Where image-based porosity is used
Areal porosity from a cross-section or surface image is a workhorse measurement across materials and life-science work:
- Coatings and paint films — void content as a quality and durability indicator
- Additive manufacturing — lack-of-fusion and gas porosity in printed parts
- Foams and porous polymers — open vs. closed void structure
- Concrete and cement — air-void and capillary porosity
- Battery electrodes — porosity governs ion transport and capacity
- Sintered metals and ceramics — residual porosity after densification
- Bone scaffolds and tissue-engineering materials — designed porosity for ingrowth
The bottom line
Porosity measures the geometric void fraction; density measures mass per volume. Image analysis measures porosity directly — a 2D area fraction that estimates the 3D void fraction under a representative-sampling assumption — and can estimate bulk density only when the solid density is known. Report porosity from several fields with its spread, calibrate before quoting pore sizes in microns, and keep the density estimate flagged as an estimate.
Frequently asked questions
- Is porosity the same as density?
- No. Density is mass per unit volume (g/cm³). Porosity is the fraction of the volume that is void space, expressed as a percent or a decimal. They are related — higher porosity lowers bulk density — but they measure different things. You can have two materials with identical porosity but very different densities if their solid phases differ.
- Can image analysis measure density directly?
- No. Image analysis measures geometry — it segments pores from solid and computes the void area fraction in the plane of the image. It does not weigh anything, so it cannot read density directly. It can estimate bulk density only when you separately know the true (solid) density of the material and the imaged section is representative of the bulk.
- How does image analysis estimate porosity?
- In three steps: segment the pores from the surrounding material (by grayscale threshold or AI), measure the total pore area inside a defined region of interest, and divide by the region area. The result is an areal porosity — the percentage of the cross-section that is void.
- Does a 2D area fraction equal the true 3D porosity?
- Only under a stereological assumption. By the Delesse principle, the expected area fraction of a phase on a random section equals its volume fraction — but that holds on average, for a representative and unbiased section. Anisotropic pore structures, a single non-representative field, or sectioning bias can make a single 2D measurement diverge from the true 3D porosity. Average several fields and report the spread.
- How do I convert porosity to a density estimate?
- Use the relationship bulk density = solid density × (1 − porosity). If a material with a solid (true) density of 1.25 g/cm³ has a measured porosity of 0.60, its estimated bulk density is 1.25 × (1 − 0.60) = 0.50 g/cm³. This is an estimate that depends on the porosity being representative and the solid density being known.
References
- ASTM E562 — Standard Test Method for Determining Volume Fraction by Systematic Manual Point Count. ASTM International, West Conshohocken, PA. View standard
- ASTM E2109 — Standard Test Methods for Determining Area Percentage Porosity in Thermal Sprayed Coatings. ASTM International, West Conshohocken, PA. View standard
- Russ, J. C., & DeHoff, R. T. (2000). Practical Stereology (2nd ed.). Springer (Plenum Press), New York. The Delesse principle: the expected areal fraction of a phase equals its volume fraction.
