Chapter 269 Xu You's Three Plans
After a pause, Ren Wendou continued:
"Of course, high resolution alone is not the goal, but we need to be able to clearly see each ground target. Ultra-high directivity and ultra-high stability are what we must ensure."
"Then what is the maximum carrying capacity of our rocket? Can the lens be made very large?"
"As long as it can be achieved with our existing rocket capacity, there is no problem."
"Teacher Ren, our Journey 9 is still unusable right now, right?"
"Zhengtu-9 is still under development and is expected to be put into use within a few years. However, it is unlikely to be available in the short term."
After receiving a series of answers from Ren Wendou, Xu You roughly understood the current situation.
Within the existing conditions, we will make every effort to develop satellites with higher resolution.
Even paying a little more is no problem.
But some things cannot be solved with money.
For example, the carrying capacity of a rocket.
If we want to solve this problem by improving the rocket's carrying capacity.
Then the problem will become more complicated.
Xu You thinks that he should try his best to think about how to achieve a higher resolution with the existing resources.
Because of the limited carrying capacity of rockets, sending a particularly large mirror at a time is definitely not a good option.
If you think from the perspective of mirror splicing or optical film assembly.
It is a feasible direction.
Just send the lenses or optical films into space several times and then assemble them.
Of course, in this case, how to assemble these lenses in space is another key issue.
After finishing the communication with Ren Wendou, Xu You began to design a specific plan.
So far, Xu You has three general directions.
The first direction is to optimize the previous generation of high-resolution satellites.
Through research, Xu You discovered that the previous generation of high-resolution satellites still has certain potential in terms of resolution, and the resolution can be further improved without changing the main hardware.
Although it won't be improved several times, there is still hope that it can be achieved within ten centimeters.
The second direction is to adopt multi-lens splicing technology.
This direction can essentially improve the resolution of satellites.
The problem is that reconnaissance satellites are relatively heavy, which poses a huge challenge to the carrying capacity of launch vehicles.
The third direction is to utilize diffractive optical film technology.
Use very thin optical films to replace ordinary optical lenses.
Because optical film is much lighter than optical lenses.
This plan does not require such .
But the problem is that it also requires assembly, and its imaging technology is not as mature as traditional optical lenses.
It can be said that each of the three methods has its own advantages and disadvantages.
It is necessary to make more reasonable choices based on different situations.
On this day, Wu Jing organized everyone and held a meeting.
At the meeting, the main focus of communication was on the design plan of the new generation of high-resolution satellites.
"Teacher Xu, please tell me about your design plan."
Xu You presented his three plans to everyone and introduced the pros and cons of each plan.
Listening to Xu You's explanation, everyone present expressed admiration for Xu You's ideas.
They didn't expect that Xu You had only been here for a few days, but he had not only completed the learning of such a huge amount of knowledge, but also had his own ideas about satellite design.
There are even some plans that even senior designers have never thought of.
"Teacher Xu, according to what you said, there is still room for improvement regarding the previous generation of high-resolution satellites?" Wu Jing asked.
“Yes, there is room for optimization. Ideally, we hope to increase its resolution to around 50mm without changing the main hardware.”
Everyone was shocked to hear this number .
Even if they used larger optical lenses, they did not have enough confidence that the satellite's ground resolution could reach such a fine level.
In this case, the resolution is more than doubled compared to the previous satellite with the highest resolution.
"This plan should be the one with the lowest cost and the fastest speed, right?"
"Yes. Of course, there is a limit to the improvement of this solution. No matter how hard we try, it is difficult to improve the resolution to within 10mm."
This is like an ordinary person who can run the 100 meters in 13 seconds. After professional training, he may be able to run it in around 12 seconds, or even faster.
But no matter how much training you do, it is difficult to break 11 seconds.
This is the limitation of the hardware, which determines the upper limit of capabilities.
If the other two solutions were used, it would not be impossible to reduce the resolution to within 10mm as long as Xu You was given enough time.
After some discussion, everyone felt that the first option could be given priority consideration.
Xu You could also see that, if possible, everyone hoped to launch such a high-resolution satellite into space as soon as possible.
"In that case, let's start with Teacher Xu's first plan. If this plan doesn't work well, we'll try other plans," said Wu Jing.
Anyway, the remaining two options will take longer to study.
Putting it at the back of the priority list won’t have much impact.
After the meeting, Xu You began to study Plan 1.
In fact, before the meeting, Xu You already had a basic plan for this plan.
According to Xu You's estimation, he is confident that after completing the optimization, the resolution can be reduced to between 30mm and 50mm.
Because this is the theoretical resolution limit of the previous generation of high-resolution satellites.
Xu You returned to his office and began his busy work.
The first thing Xu You needed to do was to optimize based on the previous generation algorithm.
Xu You entered the deep learning state, and with highly concentrated attention, he quickly entered the flow state.
Based on the knowledge he had recently learned and his previous experience, Xu You determined the optimization direction of the algorithm.
"Using the concept of depthwise separable convolution, decoupling the convolutional layers and reducing the number of model parameters can improve detection efficiency."
Xu You analyzed the data from the previous generation of high-resolution satellites and began to optimize the algorithm.
In previous algorithms, it was impossible to locate and distinguish multiple categories of targets at the same time, and they were easily affected by changes in external environment such as lighting and shooting angle, which affected the final observation results.
Therefore, only by reducing the resolution to 100mm can the ultra-high stability of satellite imaging be best guaranteed.
"Got it! Convert this target positioning and detection problem into a regression problem. In this way..."
