SHANGHAI MICROCRE OPTICS-MECH TECH CO., LTDAbout MICROCRE400-812-7833
METHODS

Quantitative microscopy: from pixels to verifiable data

Image analysis converts microstructural features into dimensions, area fractions or counts. Representative sampling, pixel calibration and segmentation are essential. A magnification label does not replace scale calibration, and one two-dimensional field cannot directly describe a specimen's three-dimensional structure. Define the rules before applying them to the complete image set.

Theory & worked calculations

Item

Translation pending review.

二维孔隙面积:从像素到面积分数:模型与量纲

Equation

Ap=Np s²;Avalid=Nvalid s²

Variables & dimensions

Translation pending review.

有效像素数=1000000 px;孔隙像素数=24000 px;比例尺s=0.5 µm/px

Assumptions & limits

Translation pending review.

本例没有给出真实三维孔隙率。分割误差、制样拔出、孔隙盖层、视场选择及截面代表性均可主导最终偏差。

Worked example

Translation pending review.

假设输入演算:24000×0.5² = 6000 µm²;1000000×0.5² = 250000 µm²;6000/250000×100 = 2.4 %。同一图像统一缩放时,像素数比仍可保持2.4%,但绝对面积和粒径需要正确校准。这个区别解释了为什么面积分数正常并不能证明尺度标定正确。

Reporting requirements

Translation pending review.

以试样编号关联原始图像、测量序列、设备配置及方法版本。

Translation pending review.

分别报告重复性、异常处理及本模型未覆盖的影响因素,不用演算示例替代实测。

Common misconceptions

Translation pending review.

数字放大不能恢复光学分辨能力;比例尺、像元采样与边界识别应分别核查。

Translation pending review.

一个试样的多个视场不是多个独立试样;二维面积结果不自动等于三维体积指标。

Workflow & steps

01

Plan sampling and imaging

Define the target features, section orientation, field-selection rule and statistics. Keep illumination, exposure and focus consistent and avoid selecting only apparently typical fields. Preserve raw images and save enhancements as separate copies.

02

Calibrate spatial scale

Image a traceable scale using the same objective, camera and acquisition settings to obtain μm/px. Recheck after changing the objective, camera adapter, pixel binning or image resizing, and assess field distortion where relevant.

03

Segment and review

Choose segmentation rules from feature intensity, colour or shape, retaining thresholds and preprocessing parameters. Inspect outline overlays for missed and false detections. Document separation of touching particles, hole filling and treatment of edge objects.

04

Calculate and test sensitivity

Area fraction is target area divided by the valid analysis area. Equivalent-circle diameter is D=√(4A/π), with calibrated area A. Compare results across a reasonable threshold range to distinguish segmentation sensitivity from actual specimen differences.

05

Report reproducibly

Summarise counts, distributions and scatter at field and specimen levels. Many particles in one specimen are not independent specimens. Export scale information, segmentation overlays, measurement tables and rule versions, stating the limits of two-dimensional section analysis.

Checks

  • • Match calibration to acquisition settings and track resizing.
  • • Save raw images, masks, thresholds and edge-object rules.
  • • State field count, specimen count and the interpretation limits of two-dimensional results.

Illustrative example

Teaching example: at 0.5 μm/px, an object covering 400 pixels has area 100 μm² and equivalent-circle diameter about 11.28 μm. This is a geometric calculation; actual measurements also depend on boundaries, resolution and scale uncertainty.

Related instruments & resources

Links support navigation and do not establish a model’s performance or standards compliance.

Your next research question. Let’s explore.

Tell us about your material, specimen and measurement goal.

Discuss your application ↗
Find your instrument
⌕
Search products, cases, solutions, standards and literature.