I’m writing like RFC, criticize, because I studied physics and mathematics for a long time, but the ideas seem very promising.

Now it is common to install a lot of cameras around the perimeter of the protected area or “fisheye” on the bumper of a car with unpleasant distortions. If you look in all directions, you won’t get enough cameras (K), especially if they are not wide-focused, but of high quality.

After all, from the middle of the object from above, you can view basically everything around and even up/down to a full sphere for a few Hz with a small “gadget” (mirror sensor, ZD) and suggest the most interesting/dangerous objects with the help of neural networks (NS), determine the distance to any, its dimensions, calculate the speed, direction and current danger. There is a job for all IT specialists! ðŸ˜‰

If you simply rotate K evenly, looking it in all directions in turn, then the blur during each exposure will be comparable to the size of the entire frame. We want to use the efficiency as much as possible, without stopping the rotation of the K and without taking breaks between shootings. And by the time we turn to a new place, everything will be completely blurred. Plus, it will be quite difficult to survey the entire sphere purely mechanically. And it is very difficult to divert wild streams of generated data to the host from the rotating C.

It looks very simple:

At the top left is a schematic view of the ZD from the side, at the bottom left is a top view, and on the right is an explanation of the physical principle for non-lubrication, because everything is continuously rotating and exposing itself in the process of movement.

We have two large coaxial monotonically rotating rings with mirrors (KZ, gray), rotating with the same angular velocity in one direction, mirrors (Z, black) are rigidly attached to them: from the outer CCZ they “look” in different directions that we need, receive light beams from there (PS, red) and reflect to the middle, to the inner CCZ, the Z of which are located more regularly and redirect the PS from the side down. to the camera (cameras, K, dark red).

Everything can be stamped from silver-painted plastic, but with perfectly flat surfaces, it is inexpensive, because the main thing here is fast K and smooth rotation. Even cheap smartphones can come up at the bottom, which over time have figured out the speed of the CCG rotation and adjusted their frame rate of shooting, and even control the rotation of the CCG on Bluetooth. As well as communicating with each other, matching their pictures for the stereo method and launching recognition NS. And everything is covered with a transparent cap for mechanical protection.

Ideally, if we want to see and evaluate a suspicious object of about 10 cm in size at a distance of 100 m at a distance of 100 m at a speed of 90 km/h through a camera with a resolution of 2K in 10 pixels for accurate recognition and representation of the level of danger (this is 1 cm in 1 pixel, sine angle 1/10000), then the angle of view K should be of the order of arcsin(0.2)â‰ˆ12Â°. If absolutely everything around is regularly inspected with such a field of view (FD), then you need to take about (90/12==7.5)*4*(7.5*2)/1.5 pictures, 300 frames. Modern cameras can be larger than Hz, but it’s still better to reasonably limit the areas of interest. If there is a sharp turn, ascent, bend of the mountain ahead, or at the intersection you need AI from reckless drivers from the side, flying at a red light, then it is better to have a 360Â° circle and a height of Â±30Â°. … In short, over time, “science” will reach a full field of view at least 10 Hz, and NS will become faster and smarter…

On the car, there is still shaking, trembling, turning of the PZ from the steering wheel, these are separate problems, in contrast to a stable CCZ on the pole. Each CCZ will have some stability due to rotation, such as a gyroscope, but in general, you need a soft suspension of the ZD relative to the place of attachment.

In principle, two CCPs can be “glued” into one by placing several prisms with two reflective faces on it in a circle, instead of the neighboring Z on a pair of CCPs. If the usual K is on the order of a few millimeters, then such a set of CCZ will take up a few centimeters in diameter, and in height – the size of the pupil K.

The speed of rotation can be any, modern technology will allow, because electric motors and car motors have a lot of Hz, and everything works for years. It all depends on the speed abilities of K. Usually, immediately after the completion of the E of the previous frame, the capture of the new one and the unloading of the past into the parallel begins, the delay is minimal. It is also necessary to be able to abruptly switch the PZ K from one prism to another, although the best way out would be to simultaneously E PS from 2 or more prisms to one large M (matrix) of one K â€“ if possible, but to different places M. Then it will be possible to regularly take frames from M and sum up. And even by neighboring ones, it is possible to identify the most dangerous fast-moving objects, but not often for the same direction, once in turn.

Let’s move on to the right side of the figure, the proof of the non-blurring of the image inside K on its light-sensitive M when the PS is reflected from a pair of rotating mirrors (mirror pair, ZP). While K is looking at the ZP flying over it, it is exposing and should get a standing image of that region on M. Another ZP will come to the PZ and E will start from the other direction.

There is a law in optics: an ordinary lens gathers all the parallel rays passing through it into one point at its focal length. Therefore, if the same PS moves a little in parallel, then it is not terrible, as long as it hits the lens, then it will be refracted and hit the same pixel in M. There is a difficulty with nearby objects â€“ when the observer is displaced, the PS changes its angle away from them. A fly, for example, at 20 cm will be unsharp, although suddenly dangerous ðŸ˜‰ But this effect can be reduced by observing the object “sideways” from the radial direction, when the PO is facing tangentially to the CCZ, or as close as possible to the tangent, so that neighboring ZPs do not interfere with its PZ. In this case, the shift in the direction of the PS during the time E will be minimal.

Instead of rotating the payroll without losing the commonality, let’s leave them in place, and change the angle of the incoming substation. We need to keep the corner at the exit. We have 2 PS â€“ red and orange. On the right-top there is their reflection from the parallel Z, everything is obvious there, the angle between the outgoing substations is exactly the same as at the entrance. If the mirrors are not parallel (bottom right), then everything is the same â€“ after the first reflection, the angle between red and orange will not change, and after the second, and no matter how many more of them there are later… In fact, if we observe many more successive Z in the nearest Z, the Z of the most recent one will come down to us unchanged, only perhaps somehow cunningly reflected in relation to a certain line.

If we have only one Z, then we can remove the second Z from the right pictures, and globally the output angle between the PSs will remain the same. I can’t see yet whether just one Z will work for us instead of a salary: it is very good for them to look sideways, and if they look up/down, they definitely need a salary. The CCZ can be stuffed with both types of prisms.

If we now turn back to the global coordinate system, we will see that in the process of any rotation of group Z, the PS coming out of it from the observed object will always be parallel to itself to the initial one.

This was for a two-dimensional case, and we have a three-dimensional case â€“ from somewhere below or above, the PS can also come in handy, and you can observe the full sphere. The 2 of our black mirror lines will represent the intersections of the Z pair with the horizontal plane. In this case, the PS vectors can be divided into 2 components â€“ horizontal and vertical. When the CCP rotates around the vertical axis, the vertical part of each piece of the PS trajectory will remain unchanged, and the horizontal part is positively considered above.

The second reflection vertically downwards will be more interesting â€“ the entire horizontal vector of the substation will be sent vertically down by the inner Z to K+M. But since we have seen horizontally the parallelism of the outgoing PS to itself during the rotation of the pair Z, this secondary downward PS will also run parallel to the primary one and will be fixed on M without blurring.

In order to provide a process of simultaneous E from at least a pair of closely spaced prisms, as suggested above, it is possible to send the PS from each prism in a slightly different direction, so that on M the focusing from adjacent prisms occurs in non-overlapping areas. In this case, the K lens will be needed twice as large, but the prisms in the CCZ will also be stuffed in the tightest way, without intervals that exclude the overlap of the PS from the solids. It will be great if the places where the PS hits from all prisms in M alternate. The very first prism leaves, the E of her PS ends, the E from the next prism continues elsewhere, the E from the 3rd prism starts to the 3rd place of the M, and there is still a 4th region in the .. That is, it is even possible to “seriously” expose 2 prisms from different directions at the same time through a large lens, as well as to ensure the change of 2 extreme ones – one leaves, and the other arrives. All 4 M regions are engaged, partial E regions are regularly unloaded and summed up with the previous ones as long as it makes sense.

If the lighting conditions change, the angular rotation speeds of the CCZ can be changed.

For real-time “stereo” and its excellent “goodies” it is possible to have pairs of Z on opposite sides of the CCZ, looking in the same direction and giving binocular vision to a pair of K with a sufficiently large base, if necessary, wider than the base of our eyes, and then compare all objects and calculate distances, sizes, speeds, directions. It is even better to have 3 K at the corners of a regular triangle, each of the faces of which will serve a sector of 120Â°, or 4 K…

Another interesting task is to reduce the number of Z’s looking at a certain level vertically. So that, for example, only four Z can observe the most capacious horizontal sector in terms of height, three – from 30Â° degrees to 50Â°, two – up to 70Â°… It will definitely take 2 CCZ, and the outer one rotates several times faster than the inner one, where the Z is rigid, and the outer one is rotating in the opposite direction, in order to be exactly “parallel” to the Z of the internal CCP and send the correct PS to it. Various pairwise combinations of the Z of the external and internal CCZ will create a lot of options for the direction of the “gaze”. In this way, we save a lot of space if we have miniature very accurate motors for the external control system, and there will be no simple numerous prisms in such a scheme.

Globally, after a dense assembly and any reassembly, each ZD will need to be aligned on some equipment and in a special environment with objects, the coordinates of which are precisely known, in order to verify the given accuracy of determining all positions to the pixel.

P.S. Substantive comments and suggestions are very welcome! Thank you for your understanding.

———-

### Acknowledgment and Usage Notice

The editorial team at TechBurst Magazine acknowledges the invaluable contribution of the author of the original article that forms the foundation of our publication. We sincerely appreciate the author’s work. All images in this publication are sourced directly from the original article, where a reference to the authorâ€™s profile is provided as well. This publication respects the authorâ€™s rights and enhances the visibility of their original work. If there are any concerns or the author wishes to discuss this matter further, we welcome an open dialogue to address potential issues and find an amicable resolution. Feel free to contact us through the ‘Contact Us’ section; the link is available in the website footer.