Skip to main content

Process steps explanation

Let's take a look at the output at each step of the process in order to better understant the whole process:

1. Load captured image​

First we need to load the image that we want to process. gls_org

2. Detect the insects:​

Then we need to detect all the insects using Grounding DINO. gls_dino

3. Segment the insects:​

We need to mask the detected insects to be able to show a more precise location for each one of them. gsl_sam

4. Dilate the mask:​

In order to make sure we are covering the whole insect we need to take the black and white mask and dilate it a little. gsl_mask gsl_mask_dilated

5. Inpaint the insects:​

Using the dilated mask, we know exactly what parts need to be inpainted and thus removed from the image in a context-aware manner. gls_org gsl_lama

6. Background subtraction:​

By getting the difference between the original and inpainted images we can see precise outlines of the insects. gsl_diff

7. Mask thresholding:​

We apply a threshold to the mask in order to transform the subtraction difference into a binary (black and white) mask that outlines well the insects. gsl_diff_mask

8. Overlay masked original on yellow background:​

Once we have the final mask, we need to apply it to the original image in order to get only the detected insects and replace the rest with a yellow background. gsl_yellow

9. Pass image through DIOPSIS algorithm.​

Finally, we can pass the yellow background image through the original algorithm. gsl_diopsis