2-class Grapevine Pest Dataset of Scaphoideus titanus and Orientus ishidae on yellow Sticky traps for Insect Detection

This dataset consists of 615 images of Scaphoideus titanus (ST) and Orientus ishidae (OI) from yellow sticky traps (YST). Among these, 150 photos, which lack target insects, have been repurposed as background images. Insect annotations comprise 1329 ST and 1506 for OI, ensuring an almost class-balanced dataset. The images were acquired through four distinct methods:

• Photos from the field;
• Images of stored YST (T = 5±1°C) and reared insects within a controlled greenhouse environment;
• Digital scans of YST collected during regular monitoring activities in the fields;
• Photos from a smart trap prototype installed in our experimental vineyard.

Structure of the dataset, showing the number of images from each data source and the corresponding class annotations.

Image source	Number of images	ST annotations	OI annotations	Number of background images
Field	18	3	101	8
Laboratory	157	473	863	8
scanned	390	853	542	84
smart-trap	50	0	0	50

We provide the yellow sticky trap images already cropped in the pre-processing stage, the corresponding enhanced datasets focusing on brightness & contrast, sharpness, and a combination of both. Finally the annotations exported in YOLO format.

Dataset structure

1. crop/
2. bright/
3. sharp/
4. bright_and_sharp/ 
5. labels/

At the time of publication, this dataset is the largest publicly available resource in the control of FD vectors. Detailed documentation, along with model benchmarking and performance results is given in an accompanying journal paper: (paper under submission).

Deployment

You can use this dataset as starting point to train your own insect detection models. Open source Python scripts to deploy the trained models can be found in our Github repository.