DJI zenmuse L1 holes in point clouds issue

PIX4Dmapper PIX4Dmapper Questions/Troubleshooting
1

I am having a problem regarding my DJI L1 point clouds.

Here’s my flight parameters:
55AGL 3m/s 85:85 overlap
AOI is planted pine trees

I’m trying to get the point clouds out of this using L1 images but its giving me holes in the middle of the forest. I’ve got GCPs and i have manually tie it but i’m not sure where i am going wrong with processing the data. I’ve tried different parameters but still giving me issues.

image
image

image
image419×567 9.48 KB

image
image
image456×720 55.7 KB

2

Hi @uav2,

Welcome to the Pix4D Community.

When dealing with point cloud generation in forested areas, it’s not uncommon to encounter issues like holes in the point cloud. This can be due to several factors, such as the density of the vegetation, the complexity of the terrain, and the quality of the images captured.

Here are a few steps and considerations that might help you improve the point cloud:

  1. Overlap : Ensure that there is sufficient frontal and side overlap in your images. This is crucial for forested areas as the dense canopy can make it difficult to find common keypoints between images. An overlap of 70-80% or higher is often recommended for such environments.

  2. Image Quality : Check the quality of your images. Low light conditions, which are common under a dense canopy, can affect the quality of the images and lead to poor feature matching.

  3. Ground Control Points GCPs : You’ve mentioned using GCPs, which is excellent. Make sure they are well-distributed throughout the project area, including the forested sections, to provide good references for the model.

  4. Manual Tie Points MTPs : Adding MTPs in the problematic areas can help improve the point cloud. Place them in areas where the automatic processing may have difficulty, such as clearings or along the edges of the forest where the canopy is less dense.

  5. Processing Parameters : Adjust the processing parameters in PIX4Dmapper. For instance, under the Point Clouds tab in the processing options, you can change the ‘Minimum Number of Matches’ to 2 instead of the default 3. This allows the software to use keypoints that are identified in at least two images, which can help fill in gaps in dense vegetation areas.

  6. Flight Altitude : Consider the altitude at which the images were captured. Flying at a higher altitude can increase the chances of capturing more of the ground through the canopy gaps, but it can also reduce the resolution of the images.

  7. Quality Report : Analyze the Quality Report generated by PIX4Dmapper. It can provide insights into potential issues with the image acquisition or processing stages.

If you’ve tried different parameters without success, it may be necessary to revisit your flight plan or even consider capturing additional images with higher overlap or from different angles to increase the chances of generating a more complete point cloud.

Remember, processing forested areas is challenging, and sometimes multiple processing iterations and adjustments are needed to achieve the desired results.

You can find additional information here:

Cheers,
Daniele