In this year’s Google I/O Conference, the Gopro and Google released a new virtual reality camera Odyssey is equipped with 16 cameras. Today, the device is finally on sale, priced at up to $ 15000. However, the device is not intended for a general audience, only “professional content producers and publishers” are eligible to purchase, purchase is required before submitting an application. Unmanned aerial vehicle competition heats up GoPro
$ 15000 price is really expensive, Google Google 10 times Glass glasses, however, is undeniable, it equipment is worth the price.
Buy a Odyssey means you will get top 16 Gopro Hero 4 Black camera, a microphone, and the necessary cables and a package (users). We from the outside view, Odyssey seems to see a circle 16 GoPro camera, in fact, Odyssey made the process more complex. Google Vice President of product ClayBavor, Odyssey made up of 3 parts: lenses, automatic integration of the software, platforms and processing materials.
Some content producers do not want to get involved in virtual reality games necessary to take the DIY way, Odyssey is the best available option.
Odyssey is designed specifically for Google Jump virtual platform camera device. Jump this year, Google announced at Google I/O Conference virtual system platform. Gopro is the first manufacturer to launch products based on the system, Jump videos will be uploaded to Youtube, the future users can use Youtube to see 360-degree panoramic video. In addition, users can also Google Cardboard and other virtual reality equipment, as well as other virtual reality device that will soon be listed.
Odyssey helps users to achieve 360 degree panoramic photography. Shoot original video after JUMP after the conversion, will generate very realistic 3D virtual reality video.
Odyssey is how to achieve a 360-degree panoramic video recording, if you’re interested, you can read on and see if it works.
Software know the relative distance between the two cameras, locations, so as long as it can calculate the pixel scene between the two cameras, you can get the depth of a point in the scene. Using these dimensional data, you can distort a picture to get a so-called “solid vertical” view: basically, each camera picture is correct parallax.
When you transition from a webcam picture to another camera captured images, slightly a little perspective. To connection picture (or established 3D model), you from near of camera in found phase match of regional, like you from a, camera found a group pixel, then in b camera also to find also of a group (this a group refers to of is with a related parts, like with a leaves, different location camera took out of picture will slightly different), adjacent camera took out of image difference not too big, than up those not calibration of camera also will great upgrade processing speed. Currently stitching software such as Videostitch or Kolor (recently acquired by GoPro) because it is a universal software will do some guessing your shooting equipment structure, but if it is calibrated camera such as Google, you only need to pass the strict mathematical algorithms.
When you find your matching region, also known distance of these regions in the picture is poor, you can calculate the camera array and the physical distance between the region. If an object, such as the sky in infinity, then out of the picture two camera parts are very close in the sky. When the object is near the camera, objects on the screen, will get larger, and two adjacent cameras out of the image will have a larger distance deviation. Concrete deviation determined entirely by distance from the camera array, so you can get a very precise object location.