Wavy artifacts and stitching issues in Mavic 3M Orthomosaic (Agriculture)

PIX4Dmapper
1

Hello Pix4D Community,

I am encountering significant distortion issues in an orthomosaic generated from DJI Mavic 3M data.

I processed the RGB imagery in a standalone project, separate from the multispectral data.

As seen in the attached image, the final orthomosaic shows severe wavy distortions and jagged edges along the crop rows. The linear features are not aligned correctly, resulting in a “smeared” or shifting effect between stitching seams.

Has anyone experienced similar issues with Mavic 3M RGB data over agricultural fields? Any recommendations on processing settings to improve the linearity and alignment of the rows would be appreciated.

Captura de pantalla 2025-12-09 095852
Captura de pantalla 2025-12-09 0958521920×1498 1010 KB

2

Hi gonzaloscarpin,

Can you attach your PDF quality report to this posting? We’ll take a look to see what might be going on with it.

3

Yes of course

RGB_report.pdf (3.3 MB)

4

This appears to be the reflectance map, rather than the orthomosaic. For the orthomosaic, color balancing will also be applied. This means that PIX4Dmapper will try to adjust the intensity of the colors in each image so that the images fit better together. The goal of this operation is to produce a more visually pleasing result. For RGB data, use the orthomosaic for better results.

5

Thanks for the clarification. You are correct—I am focusing on the Reflectance Map because my end goal is Vegetation Indices and I am having this issue working with both RGB and Multispectral images.

However, the issue remains that the geometry is heavily distorted (wavy/melted ‘lupa’ effect on the crop rows).

I suspect this is due to the DSM noise on the crop canopy being projected into the Reflectance Map.

If anyone has specific Mavic 3M processing parameters for Step 3 that balance spatial accuracy with map smoothness, I would appreciate the tips!"

6

Hello @gonzaloscarpin, These artifacts at the edge of the field row are very common while mapping vegetation due to its complex structure. You can’t completely remove it. PIX4Dmapper relies on visual similarities between overlapping images to reconstruct the model. Trees and dense vegetation, due to their complex geometry (thousands of branches and leaves), often appear very different between overlapping images. This approximation creates visual distortions and artifacts when generating the Orthomosaic or the reflectance map. The way to minimize these artifacts is to increase the flight height and also increase the image overlap. I would also recommend that you go through the support article below.

The other option would be to try PIX4Dfields, which can help minimize the distortion you are seeing. I suggest trying the product to see if it meets your requirements.