Erase Objects from Drone Footage

Table of Contents

1. Introduction
2. Challenges of Removing Objects from Footage with Camera Movement
3. Tools in Fusion to Remove Objects
4. Example 1: Removing a Tree from Stationary Footage
5. Example 2: Removing a Car from Footage with Motion
6. Using the Planar Tracker for Removing Objects
7. Steadying the Planar Track to Stabilize the Image
8. Painting Out the Object with the Paint Node
9. Creating a Clean Patch for Distorted Surfaces
10. Merging the Patched Image with the Original
11. Finalizing the Patch and Blending with the Original Image
12. Removing a Moving Object with the Planar Tracker
13. Freezing a Still Image for Painting Out the Object
14. Adjusting the Perspective of the Patched Image
15. Masking the Patch to the Road
16. Combining the Patch with the Original Image
17. Fixing Exposure Changes and Adding Noise to the Patch
18. Conclusion

Removing Objects from Footage with Camera Movement

In this article, we will explore the process of removing objects and people from footage that has camera movement, such as drone shots. We will discuss the challenges faced in removing objects from footage with camera movement and the tools available in Fusion to help with this process. We will also provide step-by-step examples of removing objects from both stationary and moving footage.

Introduction

Removing objects from footage can be a complex task, especially when dealing with footage that has camera movement. The camera's perspective is constantly changing, making it challenging to Create seamless patches that match the camera's perspective. However, Fusion provides a range of powerful tools that can help us overcome these challenges and achieve professional-looking results.

Challenges of Removing Objects from Footage with Camera Movement

The main challenge in removing objects from footage with camera movement is the constantly changing perspective of the camera. Any patch created to remove an object needs to consistently match the camera's perspective throughout the entire footage. This requires precise tracking and stabilization techniques to ensure that the removed object seamlessly blends with the surrounding area.

Tools in Fusion to Remove Objects

Fusion offers a set of tools specifically designed for removing objects from footage with camera movement. One of the key tools is the planar tracker, which helps in tracking flat surfaces and stabilizing them. By using the planar tracker, we can create accurate patches for removing objects and then match the camera's movement to maintain the original footage's perspective.

Example 1: Removing a Tree from Stationary Footage

To illustrate the process of removing objects, let's consider an example where we want to remove a stationary tree from the footage. As the tree is on a flat surface, we can use the planar tracker to track and stabilize the area. By painting out the tree and combining the patch with the original image, we can achieve a seamless result.

Example 2: Removing a Car from Footage with Motion

In the Second example, we will tackle the challenge of removing a moving car from the footage. This requires additional steps, as the surface on which the car is moving is not flat. We will still use the planar tracker to track the road, but we will need to distort the patch to match the road's contours. By using masking techniques, we can merge the patch with the original image and create a seamless result.

Using the Planar Tracker for Removing Objects

The planar tracker is a powerful tool in Fusion for removing objects from footage. It allows us to track flat surfaces and create accurate patches for removing objects. By setting reference frames and using the steady mode, we can stabilize the tracked area and paint out the object without complicated motion tracking.

Steadying the Planar Track to Stabilize the Image

Once the planar track is created, we can use the steady mode to stabilize the tracked area. This ensures that the camera's movement is matched, and we can paint out the object without affecting the overall perspective of the footage. By using the invert steady transform option, we can restore the original camera movement, resulting in a seamless patch.

Painting Out the Object with the Paint Node

To remove the object, we can use the paint node in Fusion. By selecting the clone multi-stroke tool, we can clone areas of the image to cover the object. It's essential to set the stroke duration to match the duration of the clip to ensure that the painted strokes are consistent throughout the footage. Care should be taken to Blend the strokes with the surrounding areas, considering factors like color and texture.

Creating a Clean Patch for Distorted Surfaces

Removing objects from footage with distorted surfaces requires additional steps to ensure a seamless result. In such cases, it is often helpful to create a separate patch of the area without the object. By using tools like the transform node and corner positioner, we can create a clean patch and distort it to match the contours of the surface. The merge node helps combine the patch with the original image, and a mask is used to control where the patch is applied.

Merging the Patched Image with the Original

To integrate the patched area with the original image, we can use the merge node in Fusion. By connecting the corner positioner to the merge node and adjusting the corner controls, we can Align the patch with the contours of the surface. Fine-tuning the edges of the patch by adding softness ensures a seamless blending between the patch and the original image.

Finalizing the Patch and Blending with the Original Image

By connecting the planar tracker with the merge node, we can complete the integration of the patched image with the original footage. This ensures that the patched area follows the camera's movement, providing a natural and seamless result. Any exposure changes or lack of noise in the patched area can be corrected using color corrector and grain nodes, respectively, to match the rest of the footage.

Removing a Moving Object with the Planar Tracker

Removing moving objects from footage requires specific considerations and techniques. By using the planar tracker and freezing a still image of the moving object, we can paint out the object without affecting its movement. Adjusting the perspective of the patched image using a second planar tracker helps maintain the original camera movement, ensuring a seamless result.

Conclusion

Removing objects from footage with camera movement is a challenging but achievable task using the tools available in Fusion. By understanding the principles of planar tracking, stabilization, and painting techniques, You can successfully remove unwanted objects from your footage and achieve professional results. Experiment with different scenarios and practice using the tools to expand your skills in object removal.