home / skills / a5c-ai / babysitter / tem-image-analyzer

This skill analyzes TEM images to automatically detect particles, measure sizes, classify morphology, and index crystallography for nanomaterial

npx playbooks add skill a5c-ai/babysitter --skill tem-image-analyzer

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
2.3 KB
---
name: tem-image-analyzer
description: Transmission Electron Microscopy image analysis skill for nanoparticle size, morphology, and crystallography assessment
allowed-tools:
  - Read
  - Write
  - Glob
  - Grep
  - Bash
metadata:
  specialization: nanotechnology
  domain: science
  category: microscopy-characterization
  priority: high
  phase: 6
  tools-libraries:
    - ImageJ/Fiji
    - Gatan DigitalMicrograph
    - JEMS
    - CryoSPARC
---

# TEM Image Analyzer

## Purpose

The TEM Image Analyzer skill provides comprehensive analysis of transmission electron microscopy data for nanomaterial characterization, enabling automated particle detection, size distribution analysis, and crystallographic structure determination.

## Capabilities

- Automated particle detection and sizing
- Morphology classification
- Lattice fringe analysis
- Selected area electron diffraction (SAED) indexing
- High-resolution TEM (HRTEM) analysis
- STEM-HAADF imaging

## Usage Guidelines

### Image Analysis Workflow

1. **Particle Detection**
   - Apply appropriate thresholding
   - Use watershed for touching particles
   - Count minimum 200 particles for statistics

2. **Size Measurement**
   - Calibrate pixel size from scale bar
   - Measure Feret diameter or equivalent circular diameter
   - Report mean, standard deviation, distribution

3. **Crystallographic Analysis**
   - Index SAED patterns to phase
   - Measure d-spacings from lattice fringes
   - Identify zone axis from HRTEM

## Process Integration

- Multi-Modal Nanomaterial Characterization Pipeline
- Statistical Particle Size Distribution Analysis
- In-Situ Characterization Experiment Design

## Input Schema

```json
{
  "image_path": "string",
  "analysis_type": "sizing|morphology|crystallography",
  "scale_bar": {"length": "number", "pixels": "number"},
  "expected_material": "string (for indexing)"
}
```

## Output Schema

```json
{
  "particle_statistics": {
    "count": "number",
    "mean_size": "number (nm)",
    "std_dev": "number (nm)",
    "size_distribution": {"bins": [], "counts": []}
  },
  "morphology": {
    "shapes": [{"type": "string", "fraction": "number"}],
    "aspect_ratio": "number"
  },
  "crystallography": {
    "phase": "string",
    "d_spacings": ["number (nm)"],
    "zone_axis": "string"
  }
}
```

Overview

This skill performs Transmission Electron Microscopy (TEM) image analysis focused on nanoparticle size, morphology, and crystallography. It automates particle detection and sizing, classifies morphology, and extracts crystallographic metrics from SAED and HRTEM data. The output is numerical and structured for integration into analysis pipelines and reports.

How this skill works

The skill ingests a TEM image plus optional scale-bar calibration and an analysis type (sizing, morphology, crystallography). It applies adaptive thresholding and watershed segmentation to detect particles, computes calibrated size metrics (Feret and equivalent diameters), and produces statistical summaries. For crystallography, it indexes SAED rings/patterns and measures lattice fringe spacings from HRTEM to propose phases and zone axes.

When to use it

  • Automated particle size distribution for nanoparticle synthesis batches
  • Morphology classification during materials optimization
  • Phase identification and d-spacing measurement from SAED/HRTEM
  • High-throughput screening where >200 particles are needed for statistics
  • Feeding structured TEM results into downstream modeling or reports

Best practices

  • Calibrate pixel size using an accurate scale bar before any measurements
  • Acquire images with sufficient contrast and minimal overlap; use STEM-HAADF or HRTEM for crystallographic work
  • Ensure at least ~200 particles for robust size distribution statistics
  • Provide expected material or phases to improve SAED indexing accuracy
  • Validate automated results against a small manual annotation set when possible

Example use cases

  • Quantify mean size and size distribution for a colloidal nanoparticle synthesis campaign
  • Classify particle shapes and aspect ratios to compare synthesis protocols
  • Index SAED patterns from a mixed-phase sample to determine dominant phases
  • Measure lattice fringe spacings in HRTEM to confirm crystallographic orientation and zone axis
  • Integrate per-sample outputs into a multi-modal characterization pipeline for materials R&D

FAQ

What input formats are supported?

Supply standard image files (TIFF, PNG, JPEG). Include a scale-bar calibration object for accurate sizing.

How many particles are needed for reliable statistics?

Aim for at least 200 particles per sample for meaningful size-distribution statistics and standard deviations.