lab_5
Basics of Pix4D-GCP and without GCP comparison
Introduction:
GCP (Ground Control Point) helps a lot for the land surveyors to increase their work accuracy. In the professional area, you would be asked by people, how much of accuracy in your outcome. Therefore, applying GCP points will enhance your accuracy in your UAS performance. In the previous lab exercise, we learned how to make GCP points in the study area which is Myrick Park by using GNSS rover. After that, we corrected our GCP points to be used in data processing in Pix4D. In this lab, we are going to use corrected GCP points in the Pix4D and re-run the same process what we have done in lab 3, and see if there are differences between two outcomes.
Figure 1. Study Area : Myrick Park, LaCrosse WI
Figure 2. GCP target
Before we construct the GCPs, we need to choose proper GCP target. The material of the target can be anything but should be rigid so it can be stable on the ground. The point on the middle of the target should be very correct. There should be no gaps around the center point. If there is, It would gives you margin of the error. The middle point should be observed from the images therefore, we can pick the precise points. The distribution of the GCP target would also be important. Based on the study area, your GCP targets should be assigned equally distributed. Therefore, when you are picking the GCPs in the Pix4D, it visually helps you to pick the right target points. Make sure your targets are not effected by weather conditions and shadow.
Figure 4. Holding Geo XH 6000 GPS on the center of the GCP target.
We used Geo XH 6000 GPS to record the coordinates for our GCP points. Be stable when you hold the rover and make sure it locates at the center of the target. After you finished the recording, you can now fly the mission.
Figure 5. GCPs in text file.
With your corrected GCPs that show the coordinates, you need to convert this to text file. According to the Figure 5, y is the northing, x is easting and z is geoid coordinates. If you are given this GCP text file format, you can now use this in the Pix4D for post processing.
Figure 6. Import images
Open the Pix4D software and import your images you have taken with the GCPs. The coordinate system would be same as last time.
Figure 7. Import GCPs
Open the GCP/MTP Manager and import your GCPs from your GCP text file. Make sure you choose the right order of your coordinates. For example, if your text file shows x,y,z order, you have to choose the same order in the GCP manager.
Figure 8. Picking the GCPs
Click the each labels and it shows the images with GCP target. You have to have at least 5 images with GCPs in each GCP point.
Figure 9. Picking GCP points in the image
Ensure that you zoom in the target in the image and then try to pick the point in the center of the target as precise as possible. Sometimes you might encounter the target that is hard to pick the point. This is why spotting the GCP target is important.
Figure 10. GCPs on the
After you applied the GCPs, it will appears in the mission image like Figure 10. Before you begin the processing, we need to set our processing option because, in the particular lab, we don't have plenty of time to get the outcomes. Therefore, we go to processing option, in the initial processing, in the keypoints image scale, we choose the 1/4 (quarter image size). This will save the processing time for us. More pixel size you cover, more time you will be processing.
If you are ready, you may start processing and wait for the report.
Figure 11. Project Report
As you can see the summary of the post processing, you can there is no warning sign on the georefer encing and it says there are the GCPs that has been applied in this processing and there are total 8 of them with 0.027m error. The information below shows the keypoints that were found. As you know, the more keypoints you have, the high accuracy you will get.
Figure 12. Map of Uncorrected GCP & Corrected GCP
By using differential correction in the GPS path finder, we have improved the accuracy from 3-5m to 2-3 cm. The map( Figure 12) above shows the difference between the corrected and uncorrected GCP locations. The brightgreen circle symbol indicates the corrected GCP and red triangle indicates the uncorrected GCP. As you can visually see, corrected locations are little shifted from the uncorrected locations. This may explains the correcting your GCP points is very crucial.
Figure 13. Animation of the orthomosaics ( with and without GCPs)
The animation(Figure 13) here represents the image shift in those two orthomasaics that we have made in previous lad and this lab. The first image is the without GCPs and second image is with GCPs. You can clearly see the images were shifted little bit. According to this animation, we can learn the the mosaic images are different each other after we applied the GCPs.
Study Area:
Figure 1. Study Area : Myrick Park, LaCrosse WI
Constructing GCPs:
Before we construct the GCPs, we need to choose proper GCP target. The material of the target can be anything but should be rigid so it can be stable on the ground. The point on the middle of the target should be very correct. There should be no gaps around the center point. If there is, It would gives you margin of the error. The middle point should be observed from the images therefore, we can pick the precise points. The distribution of the GCP target would also be important. Based on the study area, your GCP targets should be assigned equally distributed. Therefore, when you are picking the GCPs in the Pix4D, it visually helps you to pick the right target points. Make sure your targets are not effected by weather conditions and shadow.
Figure 3. Use wood stick to enroot the GCP target
We used Geo XH 6000 GPS to record the coordinates for our GCP points. Be stable when you hold the rover and make sure it locates at the center of the target. After you finished the recording, you can now fly the mission.
Figure 5. GCPs in text file.
With your corrected GCPs that show the coordinates, you need to convert this to text file. According to the Figure 5, y is the northing, x is easting and z is geoid coordinates. If you are given this GCP text file format, you can now use this in the Pix4D for post processing.
Pix4D post-processing:
Figure 6. Import images
Open the Pix4D software and import your images you have taken with the GCPs. The coordinate system would be same as last time.
Figure 7. Import GCPs
Open the GCP/MTP Manager and import your GCPs from your GCP text file. Make sure you choose the right order of your coordinates. For example, if your text file shows x,y,z order, you have to choose the same order in the GCP manager.
Figure 8. Picking the GCPs
Click the each labels and it shows the images with GCP target. You have to have at least 5 images with GCPs in each GCP point.
Figure 9. Picking GCP points in the image
Ensure that you zoom in the target in the image and then try to pick the point in the center of the target as precise as possible. Sometimes you might encounter the target that is hard to pick the point. This is why spotting the GCP target is important.
Figure 10. GCPs on the
After you applied the GCPs, it will appears in the mission image like Figure 10. Before you begin the processing, we need to set our processing option because, in the particular lab, we don't have plenty of time to get the outcomes. Therefore, we go to processing option, in the initial processing, in the keypoints image scale, we choose the 1/4 (quarter image size). This will save the processing time for us. More pixel size you cover, more time you will be processing.
If you are ready, you may start processing and wait for the report.
As you can see the summary of the post processing, you can there is no warning sign on the georefer encing and it says there are the GCPs that has been applied in this processing and there are total 8 of them with 0.027m error. The information below shows the keypoints that were found. As you know, the more keypoints you have, the high accuracy you will get.
Map of Uncorrected and Corrected GCP:
By using differential correction in the GPS path finder, we have improved the accuracy from 3-5m to 2-3 cm. The map( Figure 12) above shows the difference between the corrected and uncorrected GCP locations. The brightgreen circle symbol indicates the corrected GCP and red triangle indicates the uncorrected GCP. As you can visually see, corrected locations are little shifted from the uncorrected locations. This may explains the correcting your GCP points is very crucial.
Change in two Orthomosaics:
The animation(Figure 13) here represents the image shift in those two orthomasaics that we have made in previous lad and this lab. The first image is the without GCPs and second image is with GCPs. You can clearly see the images were shifted little bit. According to this animation, we can learn the the mosaic images are different each other after we applied the GCPs.
Conclusion:
The more effort you put in your work, you will get the result what you will be satisfied.
In this lab, we have learned how to use the corrected GCP points in the Pix4D. Processing the software is important but also knowing why you are using and what are the advantages and disadvantages would affect in your study is important too. In this lab, I realized that importance of the GCP and how this impacts to my performance.
Comments
Post a Comment