PhenoHUB: A Mobile Toolkit for Digital Plant Phenotyping

Overview

Field phenotyping often happens far from a lab computer, yet many routine measurements need only a phone: the angle of a leaf, the area of a plot, the weather at the site, a quick image analysis, a first look at a data table. PhenoHUB brings these tools together in one WeChat Mini Program for plant phenotyping and photosynthesis research. It runs on iOS and Android inside WeChat, needs no installation, and connects to a Python backend for the analyses that do not fit on the phone.

Modules

Phenotypic measurement

• Leaf angle. The phone's gyroscope and accelerometer give its orientation in real time. With the phone laid along the leaf blade, the reading is the leaf inclination; readings are recorded per leaf, so several leaves of a plant can be measured in sequence.
• Land area. The plot boundary is recorded as a GPS track while walking it; the enclosed polygon area is computed and converted between m², ha and mu.
• Image quantitative analysis. Plant images are processed with OpenCV to extract leaf area and colour traits, with batch processing of multiple images.

Environment

• Agricultural weather. Location-based weather for the measurement site: temperature, humidity and light conditions, from a weather service, together with agrometeorological indices.
• Location. Coordinates and altitude from GPS/BeiDou positioning, with coordinate conversion.

AI-assisted analysis

• AI chart agent. Turns a CSV file into publication-style charts in eight types: bar charts with ANOVA, heatmaps, line charts, histograms, violin plots, scatter plots and radar charts, with checks on the input data before plotting.
• AI academic assistant. Supports experimental design, choice of analysis methods and paper writing.

Data management

CSV and Excel import, data cleaning, descriptive statistics, hypothesis tests and regression, and export of results.

Architecture

Layer	Technology
Client	Native WeChat Mini Program, TDesign Miniprogram 1.8.6, LESS
Charts	Canvas API for lightweight charts, ECharts for Weixin 1.0.2 for interactive charts
Backend	Python FastAPI for image analysis, statistics and model calls

Light interaction and sensor readings stay on the phone; image analysis, statistics and AI functions run on the backend. The project is organized by tool:

PhenoHUB/
├── pages/
│   ├── hub/                         # toolbox home
│   ├── leafAngle/                   # leaf angle
│   ├── landArea/                    # land area
│   ├── agriWeather/                 # agricultural weather
│   ├── imageQuantitativeAnalysis/   # image analysis
│   ├── aiImage/                     # AI chart agent
│   ├── aiJournal/                   # AI academic assistant
│   └── my/                          # user centre
├── components/
├── utils/
└── Backend code/                    # FastAPI service

Applications

• Leaf-angle and plot-area measurement during field surveys of crop architecture.
• Recording site weather alongside phenotypic measurements.
• First-pass image and statistical analysis of experiment data in the field or greenhouse.
• Teaching mobile phenotyping methods.

Scope and limitations

Phone sensors are not survey or laboratory instruments. Leaf-angle readings depend on how the phone is aligned with the blade, and consumer GPS has metre-level error, so area measurement suits plots and fields rather than small quadrats. Weather values come from the provider's station or grid, not from sensors at the plant. Colour-based image traits are relative unless calibrated against reference measurements. Data sent to the backend and to AI models leave the phone, so sensitive data should be handled accordingly.

Access

The source code is hosted on WeChat Git with access on request. For access or collaboration, see the contact details on the CV page.