Chapter 966: Let Americans Sleep Poorly

No matter what the Indians think, the Chinese side is actually not very concerned about their pace.

Especially in the aviation industry.

It is really achieved in reality that it is better to buy than to make, and it is better to rent than to buy.

If the Indian army really uses the money invested in the LCA project to buy ready-made aircraft, the threat it can cause may be even greater.

In short, after the first batch of pre-production J-10A entered service, Chang Haonan quickly forgot what he said at the time.

During this period, his main job was to improve the image processing system of the Ocean No. 1 satellite.

Not only the satellite end, but also the part that needs to be processed on the ground after the data is transmitted back.

And the results...

Also have some clues.

January 14.

The first snow in 2002 did not seem to come later than in previous years.

Although a large number of municipal staff and people were mobilized, the heavy snow that lasted all night caused considerable trouble to road traffic early in the morning due to insufficient equipment.

So much so that when Chang Haonan got off the car, he almost stepped on empty air.

But there was no way, the date of this achievement report meeting was determined by himself, and he really couldn't blame others.

In fact, the heavy snow even caused temporary power outages in many areas.

The original meeting place suffered because of this.

Fortunately, the Fifth Academy of Aerospace Science and Technology is a key power supply target, so it can be held in another place-

The achievement report meeting is not a technical seminar, and most of the participants are leaders related to the Ocean No. 1 satellite project.

In this way, the location of the Torch Laboratory seems a bit remote.

And the Fifth Academy of Aerospace is just right.

It's a bit outrageous that the former has been working on this project for several months, but it's his first time here.

In addition, a heavy snow made everything covered with snow, and he was a little confused for a while.

Fortunately, at this time, a familiar voice sounded from not far away:

"Mr. Chang, welcome, welcome."

Chang Haonan turned around and saw Lin Senming coming to greet him with big strides:

"Mr. Lin?"

He stepped forward and shook hands with the other party:

"What a coincidence, you just arrived too?"

"Well... I saw your car coming in and found that the comrade responsible for guiding the guests had not arrived yet, so I came down from the office specially."

Lin Senming replied:

"Didn't it snow heavily last night? I was worried that I would not be able to catch up today, so I didn't go home and stayed in the office all night..."

Obviously, this person is also a person who forgets to eat and sleep when working.

So, before two sentences, the topic between the two returned to today's report meeting:

"Mr. Chang, I summarized the content I was responsible for last night, and I feel...it's difficult to come up with a sufficiently intuitive display effect."

"Although there are a lot of data and charts, those things can only be understood by experts. Today..."

He didn't continue here, but the meaning was already very clear.

Most of the people who came today were leaders. Although some of them were technical people, different fields are different. Moreover, image recognition is an extremely abstract and novel field.

However, Chang Haonan looked relaxed and even patted Lin Senming on the shoulder:

"It's okay. I'm prepared. You can organize the report meeting according to the original plan."

The latter was still a little skeptical.

But after all, Chang Haonan, who made the final decision, said so, and there was always someone to support him when the sky fell, so he nodded in the end.

...

Perhaps due to the weather, the leaders who came to the meeting generally arrived a little late.

Fortunately, it did not affect the normal start of the report meeting.

For Chang Haonan, giving a report in such a relatively closed occasion is already a piece of cake.

However, Lin Senming was still a little nervous.

After a brief opening, the first thing to do was to introduce the overall background of the satellite and the overview of software programming.

Logically speaking, these two parts are relatively bland content.

But as he listened, Chang Haonan felt something was wrong.

Too much rote learning.

Especially the part about software design.

Even if they are professional leaders, most of them don't know much about the computer field.

However, Lin Senming started with an energy functional of C-V model and LBF model, followed by a curve evolution theory and variational level set method model, and finally ended with an active contour model improved by local binary fitting...

If it is a technical meeting within the project team, this is not a big deal, but the report meeting... still needs to be concise and easy to understand.

Taking advantage of the gap between the two paragraphs, Chang Haonan looked around and found that many people on both sides of the conference table were either frowning or drowsy.

Although Ocean No. 1 is a nominal project of the ministry, even if the speech fails today, it will not cause any serious consequences.

But there will inevitably be problems with the degree of funding tilt.

More importantly, Chang Haonan is still counting on the funds from this project to continue to work on the segmentation and tracking technology of multiphase diagrams.

So, after finally getting through the first two parts, as soon as Chang Haonan took the stage, he chose to start from the middle part of the PPT.

"Dear leaders and colleagues, Comrade Lin Senming has just introduced to you the theoretical basis for our target contour segmentation. Next, let me show you our research results over the past month."

In fact, I can't completely blame Lin Senming for being boring.

If you want to show the effect of satellite image processing, first of all, you must have enough satellite images in your hand.

However, Haiyang-1 is China’s first ocean remote sensing satellite...

Therefore, when testing before, the project team used relatively abstract calculation methods to conduct testing.

But it is obviously less intuitive.

However, Chang Haonan thought of a remedy in time before the report meeting.

"Everyone, please look at these pictures."

He directly controlled the PPT to turn to the next page.

No formulas or numbers, just four pictures.

Slightly dull picture.

A dark image composed of a regular pentagon and an ellipse on a gray background.

"The color of the picture is painted on for visual effects. During the actual processing, the background pixels will be simplified to a certain extent to reduce the pressure of information transmission..."

Attendees who have just been fed too much mathematical symbols need this kind of intuitive and simple image to relieve themselves.

Chang Haonan could clearly feel that many people had cheered up again.

"Next, I will show the relationship between the evolution curve C and the target contour C* under the condition of piecewise constant value."

While speaking, Chang Haonan clicked the mouse, and a white round frame appeared on the first picture:

"In picture (a), when the evolution curve C surrounds the real target contour C*, the energy functional inside the curve C is greater than zero, while the external energy is close to zero; and in (b), when the real target contour C* surrounds When curve C is reached, the energy functional inside curve C is close to zero, and the external energy is greater than zero..."

"In other words, for any part of the evolution curve, we only need to calculate its internal and external energy functional values ​​separately to determine its positional relationship with the real target contour."

"When the calculation results of its internal and external energy are close to 0, it can be judged whether it coincides with the target contour..."

At this moment, the originally lifeless conference room suddenly became lively——

Except for some busy international waterways, there are actually not that many ships in most areas of the ocean.

In other words, most of the photos taken by ocean surveillance satellites only have a sea background.

Therefore, even if you just separate the target area and the background area and exclude data that only has background but no target, you can greatly reduce the communication pressure on the space-to-ground link and shorten the response time of the entire space-based reconnaissance system.

This does not include the efficiency of secondary identification after the ground receives the data.

Find an aircraft carrier from 10,000 photos, and find an aircraft carrier from 100 photos.

The difficulty is obviously not on the same level.

"Comrade Haonan."

The speaker was Wang Shuguang, the head of the State Oceanic Administration.

He certainly has no direct connection to the Navy.

But in principle, Haiyang-1 is a satellite belonging to the State Oceanic Administration.

Naturally, he was invited to attend today's meeting.

"I noticed that the four pictures you used all have uniform grayscale backgrounds, but the actual ocean background may not be so ideal. Even in the deep sea, there is a high probability of heterogeneous grayscales."

"Can your algorithm distinguish the grayscale differences between the backgrounds in different areas, as well as between the target to be observed and the overall background?"

This question is indeed a professional one.

It's not easy even for people in the Navy to imagine.

"sure."

Chang Haonan showed a smile:

“Actually, as long as the fitting function corresponding to the evolution curve is used to replace the gray level mean in the model in the local energy term.”

"In most cases, the picture taken in a photo does not involve too many complicated grayscale heterogeneous conditions. Combined with the weak performance of the spaceborne computer, we can basically handle it by selecting two fitting functions. Most of the time."

While talking, Chang Haonan continued to turn the pages.

This time, an overhead picture of a port appeared on the PPT.

More importantly, although the clarity is very poor, it can still be seen that there is a full-deck ship parked in the lower left corner of the picture.

"This is……"

Wang Shuguang frowned and found that things seemed not simple.

"The environment feels a bit familiar..."

"Wait a minute, isn't this Jinmen Nanjiang Wharf?"

Someone soon recognized the location:

"So this ship is the Kiev?"

"That's right."

Chang Haonan replied:

"This is a photo taken by a 244-channel CCD optical reconnaissance camera. I applied to the Air Force and got it just a few days ago."

"In addition to the Kiev, there are several other large ships whose outlines can barely be seen clearly."

"And after our processing."

He clicked the mouse again.

In an instant, except for the ground, dock buildings and the few ships, all parts involving the water surface turned into a light gray background with uniform grayscale.

“Even in shallow sea areas with complex hydrological conditions, we can still identify large surface targets!”

After finishing speaking, Chang Haonan turned his head back.

The meeting room fell silent again.

But anyone could see that this time, it was not because of listlessness.

But he was shocked and speechless for a while.

"Bang——"

Until a crisp sound broke the silence.

It was a naval leader holding a teacup, who accidentally knocked the bottom of the cup on the edge of the table.

Obviously, this technology does not even need to be put into actual use.

As long as a little bit of it is leaked, it will be enough to make Americans sleep poorly...