Applications | Hyperspectral ATH9010 for Precise Mulberry Leaf Disease and Pest Identification in Smart Sericulture

2026-09-22

An Important Factor Affecting the Sericulture Industry

Mulberry leaves are the core feed for silkworm rearing, and leaf health directly affects silkworm growth and cocoon quality. However, in actual cultivation, mulberry leaves are susceptible to diseases, pests, and environmental stress, resulting in abnormal phenomena such as leaf discolouration, spots, wilting, and tissue damage.

Diseased mulberry leaf

Traditional disease and pest detection mainly relies on manual observation, which suffers from low efficiency, strong subjectivity, and difficulty in identifying early symptoms. For some diseases and pests, before obvious symptoms appear to the naked eye, the pigments, moisture, and tissue structure inside the mulberry leaf have already changed. Therefore, how to achieve early, rapid, and non‑destructive identification of mulberry leaf diseases and pests has become an important research direction for smart sericulture cultivation.

01 Hyperspectral Imaging

From “Seeing Colour” to “Seeing Spectrum”

Hyperspectral imaging technology can simultaneously acquire spatial image information and continuous spectral information of mulberry leaves. Unlike ordinary RGB cameras that mainly record three colour channels (red, green, blue), hyperspectral devices can obtain spectral data across hundreds of contiguous narrow bands.

Hyperspectral data of mulberry leaves under different conditions

When mulberry leaves are infected by pathogens or affected by pests, their chlorophyll, moisture, cell structure, and internal chemical composition change. These changes are reflected in spectral responses at different bands.

By comparing the spectra of healthy and diseased/pest‑affected mulberry leaves, characteristic bands can be extracted. Combined with spectral preprocessing, feature selection, and machine learning algorithms, a disease and pest identification model can be established, enabling a transition from “manual leaf inspection” to “data‑driven disease and pest identification.”

02 Conducting Hyperspectral Detection of Mulberry Leaf Diseases and Pests

In a laboratory environment, the ATH9010 hyperspectral imaging system can be used to build an experimental platform for mulberry leaf disease and pest detection.

Data collection and model establishment – ATH9010

The experiment first collects hyperspectral images of healthy mulberry leaves, leaves with different diseases, and leaves with different degrees of pest damage, while recording information such as pest/disease type and severity. Subsequently, through image cropping, region of interest extraction, spectral correction, and noise processing, average spectral curves of different samples are obtained.

Further, disease/pest‑sensitive bands and spectral features are extracted, and identification models are built using algorithms such as PCA, PLSR, SVM, and Random Forest to achieve classification and identification of mulberry leaf health status, disease type, and pest severity.

Annotated samples and experimental classification results

Ultimately, the model can be applied to hyperspectral images to achieve visual marking of mulberry leaf disease and pest areas, transforming disease and pest detection from “naked‑eye observation” to “spectral data identification.”

03 From Laboratory Detection to Smart Sericulture

ATH9010 hyperspectral imaging technology can not only be used for mulberry leaf disease and pest identification but can also be further extended to scenarios such as mulberry leaf nutritional status evaluation, leaf moisture monitoring, growth analysis, and early warning of diseases and pests.

By building a complete technical workflow of “hyperspectral acquisition → feature extraction → model analysis → disease/pest identification → visualised output,” it can provide more objective data support for sericulture research, smart agriculture, and precision planting.

Annotated samples and experimental classification results

In the future, with the further integration of hyperspectral imaging, artificial intelligence, and UAV remote sensing technologies, it is expected to achieve a transition from single‑leaf detection to large‑area mulberry orchard monitoring, providing a new technological path for the digital and intelligent management of the sericulture industry.

UAV‑borne hyperspectral system