Visual recognition and tracking

MangoTree

Visual recognition and tracking [Copy link]

“At the moment the bullet leaves the gun, the wrist shakes rapidly, and the vibration can make the bullet turn. ”Do you still remember this classic line from my country's anti-Japanese drama?

It is unknown whether wrist shaking can make bullets turn, but what we can be sure of is that in reality, the United States has proposed and started to study the technology that can make bullets turn in the air. In 2014, the US Defense Advanced Research Projects Agency (DARPA) established the "Extreme Accuracy Ordnance Equipment" program (EXACTO), which aims to modify bullets so that they can automatically turn in the air. "We're trying to solve the problem of aiming error," Rice said. "This new system is about reducing aiming error and getting to the target faster."

And today we're going to talk about our real-time tracking and recognition technology - how to lock onto a certain person in a crowd. The friends made a test during which they used a simulated toy gun, a thermal imager and a development board to identify and track objects in motion in real time.

For 1080p images, the theoretical speed can reach 330 frames per second

For 720p images, the theoretical speed can reach 550 frames per second.

The lens can clearlytake portrait images through the thermal imager.

The development of this technology means that it is possible to remotely lock onto a person/object and conduct real-time tracking. To some extent, it does achieve the goal of remotely locking onto a target, thus truly realizing the idea of making bullets turn in the air. Mango Tree's newly developed visual recognition technology can remotely identify, track, and perform automatic aiming. We will still use tests to introduce this technology.

This is an object recognition and tracking technology based on FPGA. The whole process is completed in an FPGA. FPGA drives the camera to scan and shoot, and obtains the image for processing. Different objects are identified and tracked according to the configuration of parameters, shape, color, and size. Through the processing of the algorithm, the position of the image is obtained, and real-time and rapid response is achieved to achieve the purpose of tracking execution and automatic aiming.

While we are chasing the pace of technology, we are also chasing innovation...