Sugarcane Early Shoot Borer Dataset: 24,015 Field Images from India

24,015 high-resolution images of sugarcane pests were collected from fields in the Vellore district of Tamil Nadu, India. The dataset consists of four classes: Deadheart (9,782 images), Larva (7,554 images), Tillers (3,790 images), and Borehole (2,889 images). Images were captured under diverse field conditions using Android smartphones and validated by sugarcane experts to support research in agricultural AI and pest management.

Use Cases

• Train image classification models to identify Early Shoot Borer damage based on the four labeled classes.
• Develop object detection systems for pest monitoring based on high-resolution field images with diverse backgrounds.
• Benchmark model robustness using images captured under varying illumination and camera orientations as described.
• Support precision agriculture applications for targeted pest management based on visual symptom detection.

Strengths

• Contains 24,015 labeled images, providing a substantial volume for model training.
• Images represent real-world conditions with diverse illumination, backgrounds, and viewing orientations.
• Data is validated by sugarcane experts, suggesting domain relevance.
• Includes high-resolution images with multiple specific resolutions up to 4624×4624 pixels.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect geographic bias inherent to collection from a single district in Tamil Nadu, India.

Provenance

Source

Collected by S Maheshkumar and P Shunmuga Perumal from Vellore Institute of Technology under the guidance of Vellore Co-Op Sugar Mills Pvt. officials.

Collection Method

Images captured using Android smartphone devices with automatic focus and white balance in sugarcane fields.

Freshness

Last updated 2026-05-13 11:24:23; freshness should be verified.

Geography

Sugarcane fields in the Vellore district of Tamil Nadu, India.