Missing Eggs on Tray Dataset For Empty Slots Detection
Description
This dataset is designed for computer vision–based detection of missing eggs on egg trays. It supports research and development of automated inspection systems in poultry farming and smart agriculture. The dataset contains annotated images of missing eggs in trays, where eggs may be missing, enabling object detection models to identify missing positions accurately. It is suitable for training and evaluating deep learning models such as YOLOv5, YOLOv8, Faster R-CNN, and SSD. The dataset aims to facilitate automated egg tray inspection, missing eggs detection, and the development of real-time egg theft control systems. The dataset contains two compressed files of egg images and corresponding labels. The images folder contains the raw visual data in .jpg format used for training and validating the object detection model. These images capture egg trays under different conditions, including trays with many missing eggs and trays with almost complete egg placement. Each image shows an egg tray captured from a top or near-top view, with variations in egg arrangements, camera distance, and angle. The dataset is also available on Kaggle Public repository as Missing_Egg Dataset.
Files
Steps to reproduce
The dataset was collected from multiple small-scale poultry farms located in Offa town and Abuja, Nigeria. Images of egg trays were captured using an Android mobile phone under natural farm conditions. The original images were stored in HEIC format. To improve visibility of the egg trays and ensure consistency, the images were cropped to focus on the egg crates/trays and subsequently converted to JPEG (.jpg) format for ease of processing and compatibility with computer vision frameworks. Annotation was performed using makesense.ai, where bounding boxes were manually drawn around the objects of interest. Each annotation was saved in YOLO format as a .txt file with the same filename as its corresponding image. All annotation files were organized within the labels folder.