Analysis of Entomological Boxes: Insect Detection and Classification

ROI.py (Select Regions of Interest of an Image)

1. Use the Python command with the image you want to extract ROIs as an argument (e.g., "python ROI.py GxABT_single.bmp").
2. Select a ROI on the image with your mouse.
3. Press ENTER or SPACE to save the ROI.
4. Repeat steps 2 and 3 as long as needed.
5. Press ESC to quit.
6. Annotation txt file is saved in the "Annotation_txt" folder, and individual images are saved in a sub-folder in the "Individuals_images" folder.

crop.py (Reconstruct a Box from its Parts)

1. Use the Python command with: (e.g., "python crop.py GxABT_box 500 240")
   • Path to the folder containing the images of the box to be reconstructed
   • Number of pixels to crop on horizontal sides
   • Number of pixels to crop on vertical sides
2. The reconstructed box will be saved in the current directory.

parameters_chooser.py (Contrast Detection to Get the Number of Insects of an Entomological Box)

1. Use the Python command to launch the GUI: "python parameters_chooser.py".
2. Select an image to visualize with the button: "Choose image".
3. Adjust parameters with cursors.
4. (Optional) Save parameters to a CSV file with the button: "Save parameters".
5. (Optional) Load parameters from a CSV file with the button: "Load parameters".
6. Start counting for a box on parts (GxABT method) with the button: "Start counting (box parts)".
   • OR - Start counting for a single image (Africamuseum method) with the button: "Start counting (complete box)".
7. Annotation txt files are saved in the "Annotation_txt" folder, individual insects are saved in a sub-folder in the "Insects" folder, and annotated images with bounding boxes are saved in a sub-folder in the "Parts" folder (GxABT) or "Out" folder (Africamuseum). Parameters, number, width mean, and height mean are saved in the "out.csv" file.

gembloux_method_haar.py and africamuseum_method_haar.py (Cascade Detection to Get the Number of Insects of an Entomological Box)

1. Use the Python command with: (e.g., "python gembloux_method_haar.py 1.2 3 3 GxABT_box cascade.xml ")
   • Scale Factor
   • Minimum number of neighbors
   • Minimum size of detection
   • Path to the folder containing the images of the box (GxABT) or to the single image (Africamuseum)
   • Path to the cascade XML file
2. Annotation txt files are saved in the "Annotation_txt" folder, individual insects are saved in a sub-folder in the "Insects" folder, and annotated images with bounding boxes are saved in a sub-folder in the "Parts" folder (GxABT) or "Out" folder (Africamuseum). Parameters, number, width mean, and height mean are saved in the "out.csv" file.

yolo_custom.ipynb (Detection using YOLO to Get the Number of Insects of an Entomological Box)

Follow the instructions in the Jupyter notebook.

cnn.ipynb (Classification with a CNN)

Follow the instructions in the Jupyter notebook.

svm.ipynb (Classification with a SVM)

Follow the instructions in the Jupyter notebook.