Chapter 6362
The Charismatic Charlie Wade
Marven asked him,
âWhat kind of satellite monitoring are you talking about?â
Abren said, âItâs satellite photography.â
âThe accuracy of civilian satellite photography is already very high.â
âGoogleâs satellite images canât say that they can clearly take pictures of license plates on the ground,â
âBut they can definitely take pictures of words as big as the hood.â
âThis kind of satellite image is actually very practical,â
âAnd general military targets can be identified with the naked eye.â
Then Abren said, âIn fact, I was thinking on the way here that this current AI model feels like a nuclear power plant that specializes in producing computing power.â
âIts computing power is extremely abundant.â
âIt can not only supply a certain factory to produce a certain product,â
âBut also supply an industrial city, many factories of various types,â
âAnd produce a variety of products.â
âSo I hope that we will also prepare to build our own satellite image library like Google in the future,â
âSo that we can use AI to achieve image quality enhancement and intelligent retrieval,â
âWhich will be of great help to the security and interests of Cataclysmic Front in the Middle East,â
âAnd even in the future worldwide.â
âIt is also the core motivation for me to come and see this data center and meet you again this time.â
Marven nodded and said, âIf the civilian level can take a clear picture of such a large word on the hood from the satellite orbit,â
âThen the clarity is really amazing.â
âWho dug an underground bunker and where, it is estimated that it can also be seen clearly.â
âIf the data is updated quickly enough,â
âThe timeliness and practicality will be further enhanced.â
âThatâs right.â
Abren said: âBut the key to the problem is that the higher the image accuracy,â
âThe higher the requirements for the equipment,â
âThe greater the workload of collection,â
âAnd the larger the image database;â
âIf we compare the earth to a football field,â
âAnd there are 50,000 people on the football field,â
âWe take a group photo of 50,000 people,â
âAnd each person in the photo is probably just a few pixels,â
âAnd it is impossible to tell who is who.â
âIf you use this kind of photo as a basis,â
âIt is completely impossible to find someone from it.â
âEven if you know that the person you want to find is there,â
âYou are completely helpless in the face of such a photo;â
âIf we put the camera in the center of the football field,â
âDivide the football field into ten areas,â
âTake ten photos in each area and then synthesize them,â
âThe accuracy of the picture will increase tenfold,â
âBut because there are too many people,â
âEven if the accuracy is increased tenfold,â
âIt is estimated that we still canât tell who is who,â
âAnd we can only see a sea of people;â
âIf we divide 50,000 people into 100 squares,â
âTake a photo for each square, and then combine the 100 photos,â
âThe accuracy will be improved to a certain extent.â
âWith a little effort, we can probably tell whether the other person is a man or a woman,â
âBut at this time, the number of pictures in the database will reach 100;â
âIf we continue to improve the accuracy and want to ensure that the facial features of 50,000 people can be recognized in one photo,â
âThen we probably have to increase the material library to 1,000;â
âIf we use the highest precision satellite image ratio,â
âIt is roughly equivalent to us using a high-resolution telephoto lens to take N close-up photos of each of the 50,000 people to ensure that everyoneâs facial features and every pore can be clearly presented,â
âAnd everyoneâs facial close-up photos are included in a database.â
âIn this way, if we want to find a mole on the face of a certain person among the 50,000 people,â
âThe corresponding database must be in it.â
âWe just need to find a photo and find the one that covers the mole from the N photos.â
âBut finding a mole is relatively simple,â
âAfter all, the mole is on the personâs face.â
âAs long as you find the person, you can find the mole on his face.â
âBut if we want to find a specific mosquito in a stadium with 50,000 people,â
âEven if we know that the mosquito must have been captured in a certain photo,â
âIt is very difficult to find a mosquito from hundreds of thousands or even millions of photos;â
âIf we magnify 50,000 people to the entire earth,â
âThe size of this database is beyond imagination.â
âIn this case, it is extremely difficult for us to search and identify a moving target by manpower,â
âJust like looking for a needle in a haystack.â
âWe know the needle is in the sea,â
âBut we may never be able to find it in our lifetime.â
âBut if we feed all the data to the AI model and train it,â
âSo that it can find whatever we want in the shortest time,â
âThen AI can help us find anything we want in the shortest time.â
At this point, Abren added:
âIn addition, there is another area that AI is very good at,â
âWhich is image quality enhancement.â
âIf we take a photo from a distance and the photo is a little blurry when enlarged,â
âBut not blurry enough to be completely unrecognizable,â
âAI will automatically optimize the blurry picture based on its own understanding of the picture and the pixels,â
âMaking the blurry picture clear.â
âHowever, when facing an extremely large database,â