DATASET: Open RGB Imaging Workflow for Morphological and Morphometric Analysis of Fruits using AI: A Case Study on Almonds

# Almond Kernel and Shell Image Dataset (2022–2023) This dataset contains RGB images of almond kernels and shells collected during the 2022 and 2023 harvest seasons. ## 📁 Dataset Structure The dataset is organized into four main folders: Each image represents a sample (either kernel or shell). - `Kernel_almond_2022/`: Almond kernel images collected in 2022. - `Kernel_almond_2023/`: Almond kernel images collected in 2023. - `Shell_almond_2022/`: Almond shell images collected in 2022. - `Shell_almond_2023/`: Almond shell images collected in 2023. ## 🧬 Metadata Description Two metadata files are provided to accompany the image collection, each containing detailed information on the corresponding almond samples. ### `almond_metadata_1.csv` – Morphometric Features from Image Analysis This file includes sample-level descriptors extracted from RGB images of individual fruits (kernels or shells): - **Picture**: Name of the image file. - **Fruit**: Type of fruit (e.g., kernel or shell). - **ID**: Unique identifier for each sample. - **Year**: Year of harvest (2022 or 2023). - **Population**: Breeding population or genetic group. - **Parent_1**, **Parent_2**: Parental genotype codes. - **Fruit_number**: Within-plant fruit number or replicate. - **Length**, **Width**, **Width_25**, **Width_50**, **Width_75**: Morphometric measurements (mm). - **Area**, **Perimeter**, **Hull_area**: Geometric descriptors (mm² or mm). - **Solidity**, **Circularity**, **Ellipse_Ratio**, **Aspect_ratio**: Shape descriptors. - **L**, **a**, **b**: CIELAB color components. - **Symmetry_v**, **Symmetry_h**, **Shoulder_symmetry**: Symmetry metrics. - **PB_PC1** to **PB_PC10**: Principal components pixel-based traits. - **EF_PC1** to **EF_PC6**: Principal components from EFourier descriptors. ### `almond_metadata_2.csv` – Physical Measurements and Weight Ratios This file contains physical measurements and mean values for almond kernels and shells: - **ID**: Unique identifier for each sample (matching `almond_metadata_1.csv`). - **Year**: Year of harvest. - **Kernel_Width**: Width of the kernel (mm). - **Kernel_Thickness**: Thickness of the kernel (mm). - **Kernel_Globosity**: Sphericity index of the kernel. - **Kernel_Weight_mean**: Mean kernel weight (g). - **Shell_Weight_mean**: Mean shell weight (g). - **Ratio_weight_kernel_shell**: Kernel-to-shell weight ratio. Both files can be linked through the shared **ID** field. Gaps in the data are related to the unavailability of the data. In the case of the CIELAB color components, values from certain images were removed due to detected issues with lighting and proper functioning.