Algorithm system

Looking forward to the future1The next day, Lin Yuan was woken up by his alarm clock.

Inside the cramped, narrow rented room, there was only a bed, a table, and a chair, with no other furnishings.

Clothes were casually placed in an open suitcase, and what was most visible were various books stacked up.

“C Language Programming”

“Embedded Hardware and Software Design”

“Java Network Programming”

“Introduction to Machine Learning and Practice (Python)”

“Introduction to Motor Control”

“Fundamentals of Electromechanical Principles”

“PCB Design Outline”

“Principles of Digital Image”

.

.

There were so many of these books that they couldn’t fit on the table, and some had to be placed on the floor.

The covers of some books clearly showed faded strokes, indicating they were second-hand books Lin Yuan had acquired.

Other books had uneven layouts and rough binding, which were clearly pirated copies.

These more than 30 books, covering a wide range of knowledge in computer software, hardware, and automatic control, naturally couldn’t all be read within half a year.

Lin Yuan not only selectively studied them but also used them as a reference database.

As for why he used physical books instead of e-books...

Because studying and reading is not like reading a novel; the experience of physical books is simply incomparable to e-books.

E-books cannot achieve the tactile sensation of being able to flip through pages and take notes at any time, that feeling of unity between person and book.

Besides the space for books, there was actually a lot of empty space on this table.

However, Lin Yuan only used a very small area for daily meals or coding, because half of the table was covered with a green anti-static mat.

On top of the mat was a constant temperature soldering station, two component bins full of various small compartments, and various items like tweezers, solder wire, and circuit boards.

This was a small manual soldering workstation.

Its function was, naturally, for manually soldering PCB boards.

The computing power system guided Lin Yuan to enroll in the Computer Science Department of Zijin Aviation School.

Generally speaking, computer majors rarely have direct contact with hardware.

However, because it involved the AI direction, and AI is a comprehensive discipline, it naturally became inseparable from various motors and circuits.

Lin Yuan originally thought that these things, which were beyond the scope of traditional computer knowledge, would be useful during the interview.

But he didn’t know that the other party didn’t recognize things that were tinkered with alone in a small room.

Fortunately, he ultimately relied on the formidable power of the computing power system to achieve a “one-move victory.”

Although his admission had already received Professor Zhang’s almost certain approval, the system panel still hadn’t settled the reward.

“Task: Enroll in the Computer Science Department of Zijin Aeronautics and Astronautics University. Completion: 99.9%.”

Lin Yuan didn’t know how the computing power system calculated task completion.

He tentatively assumed that the missing 0.1 percentage point was because he hadn’t yet received the official admission letter.

With the current task temporarily concluded, and no new tasks before him, Lin Yuan could only continue his original routine—first, to continue studying the books recommended by the system, and second, to continue delivering food to support himself.

Lin Yuan glanced at the bright yellow helmet on the floor in the corner of the room; this job was really not a good one.

The weather in June was already starting to get hot, and by July and August, the air in Purple Gold, this great furnace, would feel as if it were boiling hot.

Lin Yuan’s head already ached just thinking about it.

But what else could he do if he didn’t do this?

Lin Yuan pondered: He wondered when the graduate admission letter from Zijin Aviation School would arrive.

If he had that, he could use it to get a tutoring job.

As a prospective master’s student from Zijin’s top three universities, finding a tutoring job wouldn’t be difficult.

Then he could sit in an air-conditioned room and earn money by just talking.

Helplessly, Lin Yuan still had to pick up his yellow “crown.”

He opened the door, then held that position for about ten seconds.

Lin Yuan ultimately didn’t step out.

‘I can’t be constrained by an admission letter, can I? It’s just a piece of paper.

With or without it, I will be a master’s student at Zijin Aviation School.

How about...’

Lin Yuan sat down in front of his computer and began searching using a search engine—image search: Zijin Aviation School Computer Science Department graduate admission letter.

Soon, images of a blue-backed master’s admission letter came into view.

As expected of an aviation school, even the admission letter was the color of the blue sky.

Lin Yuan couldn’t help but fantasize about the moment he received his admission letter; just thinking about it made him excited.

However, at this moment, he first needed to manually create one.

The summer vacation was approaching quickly, and the golden season for tutoring was coming.

Lin Yuan had already mentally completed his tutoring resume; all that was missing was this admission letter.

An admission letter is not like printing money, Lin Yuan thought, so he should be able to get it printed at any advertising company.

It’s just a color-printed hard card, isn’t it?

The problem was that the electronic version of the admission letter was hard to get.

Many related images could be found online, and it was easy to change the name and photo on those images using Photoshop.

But the problem was, if he wanted to pursue realism, the color and size of the admission letter still had to be as accurate as possible.

However, it was not a simple matter to obtain the color codes and dimensions of various parts of the admission letter from images retrieved online.

Firstly, even for the same admission letter, different people using different phone cameras at different times would capture different color codes.

Secondly, it would also take some effort to restore the actual dimensions of the admission letter from the images.

Lin Yuan then opened PyCharm, preparing to use Python to invoke a web crawler script.

“Learn math, physics, and chemistry well, and you can go anywhere in the world without fear.”

See, it came in handy at this moment.

To solve the color code problem, Lin Yuan planned to use a web crawler script to batch-crawl images of Zijin Aviation School graduate admission letters online.

Then, after comprehensive compilation, he would filter and take an average.

Filtering was necessary.

Because in this day and age, using beauty filters for photos is too common, and who knows how distorted the actual color codes might become.

As for what “color code” is: as the name suggests, it is the identification number for each color, equivalent to a color’s ID card.

Since all data in the computer world is digital, colors are naturally no exception.

The most common color definition format is RGB three primary colors: R (red), G (green), B (blue) three primary colors can mix to produce various different colors.

Of course, besides these, there are also CMYK, HSB, and so on.

But in the computer world, the RGB format is more widely used.

This level of web crawler is not difficult to write.

Lin Yuan only needed to crawl the search results of a certain search engine, then extract images from them, and then filter them through OCR character recognition, selecting images with the words “Zijin Aviation School Graduate Admission Letter.”

Next, perform object recognition—extract the area of the admission letter in the image.

Then create a digital image histogram, which is to statistically count each color code in the admission letter image.

After that, arrange the statistical results.

It’s like scoring in a diving competition: remove the highest and lowest scores, and take the median.

This method is relatively simple; it is essentially filtering out the median value.

However, this process still involved some content related to computers and digital image processing.

Fortunately, in this era, thanks to the powerful impetus of Moore’s Law, the ceiling of computing performance has been repeatedly broken, and computer science and AI technology have developed rapidly.

Image object recognition, which could only be attempted in large laboratories 20 years ago, can now be run on an antique-level T440.

Of course, this does not mean that image object recognition itself is simple, but rather that countless people’s open-source contributions have made this technology simple to use.