Quotation update date: 2020/10/13

Upgrade 1: Graphics card RTX20 series upgrade to RTX3080/RTX3090 series, with significant performance improvement

Upgrade 2: Desktop and cluster CPU heat pipe heat dissipation upgrade to industrial-grade water-cooled heat dissipation, with higher frequency increase

Upgrade 3 Upgrade related models desktop and cluster

Main content
(1) Desktop image processing workstation configuration scheme
(2) Portable image processing workstation configuration scheme
(3) Tilt photography image processing multi-machine cluster configuration scheme
(4) Mini mobile image processing cluster configuration scheme

In the surveying and mapping industry, the aerial image application of drones is becoming more and more extensive, and the amount of data is becoming increasingly large. The market needs computing and processing equipment from desktop to mobile portable, multi-machine cluster , mobile multi-machine cluster, etc.

Since Xi’an Kunlong Computer Company has focused on the application of aerial image processing of drones in 2013, we have accumulated a large number of performance test data of various specifications and sizes of hardware in different eras, as well as many years of rich successful cases, experienced test engineers and technical support engineers, and many years of experience in hardware configuration and performance optimization.

Entering 2020, for the latest listed computing architecture, tilt photography modeling, orthophoto processing and other applications, we have once again launched a variety of solutions with better configuration, faster speed and rich models – desktop workstations , mobile workstations, multi-machine clusters, mobile clusters, etc., with significant technical features:

• The recommended machine hardware configuration, its CPU+GPU performance, memory, hard disk io, network port, and maximum performance in every link of the computing process, without bottlenecks and no blind spots.
• A complete and reliable computing solution is provided for different computing scales and different budgets.
• Ensure that different image processing software (such as: Context Capture, Pix4D, Inpho, DJI , etc.) can achieve the ultimate performance of workstation processing

(I). Recommended hardware configuration for desktop (stand-alone) image processing workstations

Related model: H390

Air three processing, three-dimensional modeling, etc. for tilt photography/ orthography based on office silent environment

Technical points:

Adopt Intel’s 10th generation ultra-high frequency processor (10 cores 5.2GHz)

Maximum memory capacity 128GB DDR4 2933

Configure RTX graphics card with Turing architecture

Configure 2.5G Ethernet port

All-round optimization of the entire machine performance

In short, this is the fastest desktop image processing workstation on the market (2nd quarter of 2020),

Data processing data reference:

Machine configuration: 10 cores 5.2GHz + 64GB + RTX2080ti 11GB

Data volume: 3860 pictures, resolution 6000×4000,

Single machine processing capability: Air three processing time 1:10, 3D modeling time 46:9, average 82 pictures/hour

Provide remote testing verification, the above data

Recommended hardware configuration of desktop image processing workstations

(II) Tilt photography modeling hardware configuration solution (mobile portable workstation)

This is the fastest portable mobile computing architecture on the market, with the highest computing processing capacity of 10 cores 5.1GHz, a maximum memory configuration of 128GB memory, and the RTX graphics card equipped with the latest Turing architecture (maximum 2 blocks)
is positioned in the air three processing, three-dimensional modeling, and orthophoto processing of mobile tilt photography . It is much higher than the computing and storage capabilities of industrial control computers and laptops on the market, while ensuring absolute high mobile and high performance.

2.1 Recommended single-screen portable configuration

2.2 Three-screen portable workstation configuration solution

(III) Recommended multi-machine cluster configuration

The calculation volume of massive aerial image data is huge, and the processing time of a single machine’s computing power is too long and the efficiency is low. With the help of the most advanced multi-machine parallel cluster processing system, air three processing and three-dimensional modeling are carried out, which greatly shortens the processing time.

Related models:

(1) Image processing workstation (computing node)

Model: UltraLAB M390 (the latest model launched in 2020)
Technical features:
adopt Intel’s 10th generation ultra-high frequency processor, equipped with industrial-grade water-cooled radiator, Nvidia Turing architecture GPU computing card, network port is equipped with 2.5G, and hard disk is M2.SSD, ensuring stronger hardware configuration calculation, higher io bandwidth, and comprehensive optimization of the entire machine performance, ensuring that aerial image data processing is completed in the shortest time.

(2) Main control and storage server ( storage node )

Related models: UltraLAB N630i (8 disks), N630 (16 disks)
Technical features:
CPU adopts Intel Xeon second-generation scalable processor
cache disk : equipped with M2.SSD to ensure high-speed read and write (read and write 3G/s or above) and ultra-low io latency,
equipped with parallel storage: data backup is safe and reliable, with maximum capacity to 112TB (8 disks)/240TB (16 disks)
network port: equipped with dual 10G optical ports, theoretical io bandwidth reaches 2.5GB/s
, and the entire machine performance is fully optimized to ensure the best aerial image data io.

This is the latest and fastest multi-machine parallel cluster solution on the market currently (June 2020)

Solution 1 Recommended high-speed drone image processing cluster configuration (5 computing nodes)

Solution 2 Recommended high-speed drone image processing cluster configuration (10 computing nodes)

Solution 3 Recommended high-speed drone image processing cluster configuration (20 computing nodes)

Solution 4 Recommended high-speed drone image processing cluster configuration (30 computing nodes)

Solution 5 Recommended high-speed drone image processing cluster configuration (60 computing nodes)

Solution 6 Recommended high-speed drone image processing cluster configuration (90 computing nodes)