DJI has just come out with the Phantom 4, but what makes it different from the previous models?
For the past 3 years, DJI has been leading the drone industry. Their consistent innovation of the Phantom product line has allowed an estimated 7,000,000 photos to be taken, and over 12,500,000 miles to be flown collectively by DJI users around the world.
The Phantom 4 was designed to have the same iconic Phantom profile that we all know and are comfortable with, but the new design has had major changes made to make it look more streamlined and reduce drag. The motors have also been raised over the shell to increase cooling an efficiency while also giving more room for the new navigation lights.
To compensate for the larger battery, the camera has been moved farther forward which also reduces the likelihood of having the propellers in your shots. What’s really impressive is how DJI was able to seamlessly integrate the anti-vibration mount and camera module into the main shell, while keeping the design relatively slim looking.
Speaking of vibration, the propellers can now be quickly attached and released thanks to the new push and lock propeller design, taken right from the expensive Inspire 1.
Fast And Stable Flight
If you didn’t already know, the Phantom 4 now has five cameras. There’s the main camera which is used for shooting 4K video, two cameras on the front and two more on the bottom. These extra cameras are used for a few things, but the first is VPS which we saw on the Phantom 3 and Inspire 1.
The VPS (visual positioning system) on the Phantom 4 has been greatly improved do to the extra cameras. The Horizontal stability has been increased by 5x and the VPS is now capable of operating at altitudes of up to 33 feet (over 3 times higher than the Phantom 3). What all this means is that in most cases, your Phantom 4 will always be good at holding its position (even without the GPS).
There’s also a new sport mode for flying at speeds up to 45mph. In this mode, Obstacle avoidance is disabled and the max tilt angle has been increased to 45 degrees.
Flight Time And Redundancy
Like the older Phantom models, the Phantom 4 still relies on core sensor information from the IMU (inertial measurement unit), however it now has two IMU’s for better redundancy and improved accuracy. This means that even if one system fails, there will be a backup to get you back on the ground safely. The flight control algorithm has also been redesigned to improved power management and make better use of all the data coming in through the onboard sensors.
Between all of these optimizations and a larger 5350mah battery, the flight time has been increased by about 25%. So realistically, if you get 18 minutes on your Phantom 3, you should get 22.5 minutes on the Phantom 4 in the same conditions.
One of the biggest features of the Phantom 4 is the new always-on obstacle detection and avoidance. With all of the new image processing hardware onboard, the Phantom 4 uses the imaging data from all 4 cameras to build a 3D volumetric map of it’s environment. In other words, it will use two cameras at a time to see the world in the same way that we humans do, creating a sense of depth perception.
By using all of this information, the Phantom 4 will automatically detect obstacles at up to 50 feet away and intelligently make the decision to stop, or go around the obstacle.
The next big feature is Tap Fly. With this feature, you can simply tap on the screen where you want the Phantom 4 to go and it will automatically make its way to that point while avoiding any obstacles in the way. This isn’t just a way of moving the drone to a specific point though.
By continuously tapping on different locations, the Phantom 4 will automatically update its flight path and make a smooth transition to its new destination while also avoiding obstacles. What this means is that you can use it to create dynamic and smooth moving shots without actually touching the controls.
The coolest feature of the new Phantom 4 is called Active Track. This feature allows the Phantom 4 to automatically track and follow subjects (without the use of any additional hardware). All you have to do is use your finger to create a box around the subject and tap the go button. Besides the fact that this could dramatically help pilots film subjects, the way that this feature works is actually quite amazing.
With traditional image tracking algorithms, they only track images using two dimensional tracking data, and once the subject rotates (changing how it looks to the eyes of the computer) the system will get confused and lose track of the target. The way DJI’s system works is different. It will detect changes in the subjects rotation, and automatically learn what it looks like from every angle, essentially building a 3D representation of the subject you want to track.
On top of that, it uses machine learning algorithms to constantly learn what the subject looks like. So if you’re filming yourself and you happen to sit down, get on your bike or do something that changes how you look, it’s constantly relearning what you look like so it doesn’t lose track of you.
What I think?