WHY THE CAMERA NEED THE “Synchronization control”
We all know that during the flight, the drone will give a trigger-signal to the five lenses of the oblique camera. The five lenses should theoretically be exposed in absolute synchronization, and then record one POS information simultaneously. But in the actual operation process, we found that after the drone sent a trigger signal, the five lenses could not be exposed simultaneously. Why did this happen?
After the flight, we will find that the total capacity of the photos collected by different lenses is generally different. This is because when using the same compression algorithm, the complexity of ground texture features affects the data size of photos,and it will affect the camera's exposure synchronization.
Different texture features
The more complex the texture of the features, the larger the amount of data that the camera need to solve, compress, and write-in.,the more time it takes to complete these steps. If the storage-time reaches the critical point, the camera cannot respond to the shutter signal in time, and the exposure-action lags.
If the interval-time between two exposure is shorter than the time required for the camera to complete the photo cycle, the camera will miss-taken photos because it cannot complete the exposure in time. Therefore, in the course of the operation, the camera synchronization control technology must be used to unify the camera's exposure-action.
R&D of synchronization control technology
Earlier we found that After the AT in the software,the position-error of the five lenses in the air can sometimes be very large, and the position difference between the cameras can actually reach 60 ~ 100cm!
However, when we tested on the ground, we found that the synchronization of the camera is still relatively high, and the response is very timely. The R & D personnel are very confused, why is the attitude and position error of the AT solution so large?
In order to find out the reasons, at the beginning of the development of DG4pros, we added a feedback timer to the DG4pros camera to record the time difference between the drone trigger signal and the camera exposure. And tested in the following four scenarios.
Scene A: Same color and texture
Scene A: Same color and texture
Scene C:Same color,different textures
Scene D:different colors and textures
Test result statistics table
For scenes with rich colors, the time required for the camera to do Bayer calculation and write-in will increase; while for scenes with many lines, the image high-frequency information is too much, and the time required for the camera to compress will also increase.
It can be seen that if the camera sampling frequency is low and the texture is simple, the camera response is good in time; but when the camera sampling frequency is high and the texture is complex, the camera response time-difference will increase greatly. And as the frequency of taking pictures is further increased, the camera will eventually miss-taken photos.
Principle of camera synchronization control
In response to the above problems, Rainpoo added a feedback control system to the camera in order to improve the synchronization of the five lenses.
The system can measure the time- difference "T" between the drone sends the trigger signal and the exposure time of each lens. If the time difference "T" of the five lenses is within an allowable range, we think that the five lenses are working synchronously. If a certain feedback value of the five lenses is greater than the standard value, the control unit will determine that the camera has a big-time-difference, and at the next exposure, the lens will be compensated according to the difference, and finally the five lenses will exposure synchronously and the time-difference will always within the standard range.
Application of synchronization control in PPK
After controlling the synchronization of the camera, in the surveying and mapping project, PPK can be used to reduce the number of control points. At present, there are three connection methods for oblique camera and PPK :
|1||One of the five lenses is linked to PPK|
|2||All five lenses are connected to PPK|
|3||Use camera synchronization control technology to feed back the average value to PPK|
Each of the three options has advantages and disadvantages:
|1||The advantage is simple, the disadvantage is that PPK only represents the spatial position of one-lens. If the five lenses are not synchronized, it will cause the position error of other lenses to be relatively large.|
|2||The advantage is also simple, the positioning is accurate, the disadvantage is that it can only target specific differential modules|
|3||The advantages are accurate positioning, high versatility, and support for various types of differential modules. The disadvantage is that the control is more complicated and the cost is relatively higher.|
There is currently a drone using a 100HZ RTK / PPK board. The board is equipped with an Ortho camera to achieve 1: 500 topographic map control-point-free, but this technology cannot achieve absolute control-point-free for oblique photography. Because the synchronization error of the five lenses themselves is greater than the positioning accuracy of the differential, so if there is no high-synchronization oblique camera, the high-frequency difference is meaningless……
At present, this control method is passive control, and compensation will only be done after the camera synchronization error is greater than the logical threshold. Therefore, for scenes with large changes in texture, there will definitely be individual point errors greater than the threshold,. In the next generation of Rie series products, Rainpoo has developed a new control method. Compared with the current control method, the camera synchronization accuracy can be improved by at least an order of magnitude and reach ns level!