G10T-AI

Main features

1.incline45°a，small volume，light weight

Triaxial pod designed specifically for fixed wing，spherical sum45High youree of design effectively reduces wind resistance（As shown below），Three-axis stabilization effectively solves the image rotation on the two-axis pod

question，Excellent algorithm and structural design，Can effectively ensure high-speed cruising（maximum50m/s）Lower stabilization accuracy。

2.Three-axis mechanical stabilization，INideo is always horizontal，No black borders，No image spin

Traditional fixed-wing pods have to withstand the impact of faster airflow，Both are multi-frame and two-axis structures，Therefore, the image rotation of the video cannot be stabilized by the mechanical structure.

Certainly，This causes the video torotate and shake around thecenter of the screen.。Even if some pods are equipped with electronic image stabilization，Still results in a black frame on the screen、Not stable enough

Certainly，Causing operator visual fatigue。Two-axis pod especially when looking vertically downward，will withpletely lose stability。This pod is a three-axis structure

structure，Ability to perform withplete three-axis stabilization of the camera，and keep the horizon at the level of the field of view，In line with human eye habits，Supports electronic image stabilization at a lotphoto(OUR)

Function，even in strong wind、high speed、Under extreme conditions such as a lotphoto，Still keep the video stable。Dozens of hours of raw video and OSD separate recording，

Can synthesize OSD，You can also keep the original video。 

3.Electronic image stabilization at a lotphoto，even in strong wind、Under conditions such as high speed，Still keep the video stable。

4.Dozens of hours of separate recording of raw video and OSD，Can synthesize OSD，INideos can also be saved。

5.Provide multiple data and docking protocols for flight control，Complete the matching of the entire avionics system。

6.Onboard AI system

6.1Active target recognition and frame selection

Can intelligently identify people、specific targets such as cars，Supports intelligent identification of multiple targets。Intelligent analysis system based on drone images。Drone real-time images

Direct input from the pod to the airborne terminal，No need to return to server for processing，Processing and analysis in real time directly onboard，Complete identification of specific goals

Detection，And superimpose the results of recognition statistics directly on the video，Send back to superior withmandcenter，Carry out further confirmation and record。 

6.2Number density statistics

The system uses deep learning convolutional neural network technology，Extract features from video images，and determine the corresponding target category。Multi-threading is also used

processing method，Maximize and rationally allocate onboard withputing resources，Distribute image preprocessing work among multiple threads，Shorten a single frame of a video image

Getting ready for neural networks(preprocessing)time，Thereby improving the processing efficiency of the overall recognition system for video streams.，The core advantage of this system is that it

multi-thread processing mechanism，Improved system processing efficiency for drone video streams，Achieve fast and real-time detection results for small targets，

Provide law enforcement officers with smooth and clear identification results。

6.3AR geographical information overlay

AR overlay of predicted targets and real-time video，Such as real-time display of pipelines、Names of cities and buildings, etc.。Standard kml data can be downloaded，Can be carried out according to

Overlay of kml data required by the protocol。Method to realize：Memory card，Customer input，Be private。

Pod parameter information

Basic Information

product name：Single light pod

Product number：G10T-30X-AI

Basic parameters

weight：530g

size：high154mm

Power consumption：conventional≤11.5IN，peak≤43IN

PTZ characteristics

type：Three-axis mechanical stabilization

Pitch：+45°～-135°

course：360°*N

Roll mechanical limit：±60°

Stable accuracy：≤0.02°

INisible light camera lens

30X optical zoom           

f=4.3mm(wide)～129.0mm(a lot）           

F1.6～F4.7

sensor：Sony Starlight CMOS

Sensor image format：FHD,1080p@30

FOIN：63.7°(wide angle end)to2.3°（Telephoto end）

Electronic zoom：2times，30times optical，comprehensive60x zoom capability

environmental adaptability

Operating temperature：-20°C～55°C

Tracking module

Tracking module：External

track target：48*48～384*384pixel*pixel

Track algorithm time：＜20ms/frame

Detect target：48*48～384*384pixel*pixel

Detection algorithm time：30ms/frame

Encoded video output：Support encoded video output，Ethernet，SupportH265H264coding，Default isH265coding

picture within picture：Support picture-in-picture

OSD:INith OSD function，Characters can be displayed superimposed，pod angle、detection frame、tracking window、point north、device status、Target coordinates

external storage：64G

power supply：18～28IN

Power consumption：standby≤10IN，Turn on tracking+AI power consumption≤22.26IN

Size power consumption：115×102×34.8mm, ≤420g（Includes fan）

interface

interface video interface：Network port、Synchronize422（customizable）、HDMI-OUT

pod control interface：CAN、S-BUS、RS232、TTL、asynchronous422（customizable） 

The above is-AI version，In addition there are：-HS normal version（The pod body is the same，The tracking module is different）

interface video interface：SDI、HDMI

pod control interface：R232、CAN、S-BUS