360 °レンズ
個数
1
select
select
カートに追加 カートに追加
テスト要請テスト要請

TCPW360-48 Multi マルチミラーシステム

詳細情報
Ai1は シングルショットを使用して得られる複数のビューからのオブジェクトの検査および品質管理のためのトータルビジョンソリューションです。 •3つの異なるFOVと3つの異なる角度で利用できます。 •ユーザーフレンドリーなインターフェースに総合されあたソフトウェア •プラグアンドプレイのセットアップにより、時間を大幅に節約できます。 •簡単なハードウェア及びソフトウェアのテクニカルサポート. •簡単な4つステッププロセス: -獲得 -ラベリング -クラウドでのトレーニング -強力な総合エッジデバイスでのインファレンスの実行時価
Brand : Far Island
Fov : 46x38.5 mm
Mag : 0.184 x
DOF : 20

SmartPC-Ai1 シリーズ

Our main customers requested a new device that could solve common problems in the Industrial Computer Vision field: durability, simplicity of deployment, ultra-compact design, and top performance.

So we designed the SmartPC-A11 to solve the challenges we face in Industrial Computer Vision.

A Total, All-in-One, Plug & Play Vision Solution that integrates optical systems, deep-learning software, illuminations, and a powerful AI edge device into a simple, unified form factor. It means no more hundreds of different cables to choose and connect, no more different vendors for different parts of the vision system, no more dozens of different failure points to analyse in case of an incident.

At the heart of the SmartPC-Ai1 Series, there is the SmartPC-Ai1 Standalone, an edge device with highly optimized built-in software, enhancing operational efficiency in edge computing. It can transform any system into a state-of-the-art deep learning-based solution in a few minutes. The optical system is from famous Italian brand Opto Engineering, while the computing power comes from NVIDIA AI-Edge devices (8-cores and 12-cores). The high-power illumination is handled by an internal controller and totally managed by onboard software.

Another innovation that we introduced is the automated remote cloud training. Training task for Deep Learning is a big resource and time-consuming task. The traditional approach solves this problem with very expensive workstations that are good for training but overpower all the other tasks. We removed this requirement from the on-site production device, which now can be robust, fast for inference and low cost. After the labeling, the dataset is transferred to the cloud in a click of a button and the training starts automatically. When the training is completed by our powerful servers, the inference-model is transferred back to the production device, and the operations can start.

Concept

Traditional Approach

Problems

  • The integration requires many different skills and expertise.
  • The BOM contains dozens of components.
  • The assembly and maintenance cost are high (time consuming).
  • The software integration has too many levels of complexities (computer vision, algorithms, network, peripherals communication, drivers' compatibilities and so on)
Far Island Corporation Ltd
Seoil bld. 703, Sapyeong Dae-ro 353, Seocho Gu, Seoul, Korea.

+82 10 396 86098
info@far-island.com
Opto EngineeringRicohTms Lite
Distributor