By utilizing Houmo HaloDrive ®️ 30 chip, the edge computing power level is redefined, and the edge computing power is increased to 60TOPS, thus achieving the “local collection and local computing” of front-end video data.
For complex industrial applications, we provide superior customized development and deployment services, and conduct the underlying optimization design from the network architecture and operator level of the algorithm model. The algorithm is hardware-based to achieve superior hardware acceleration.
Starting from the massive video data in the industrial field, the video analysis concept of “real-time, full volume and whole network” is put forward to enhance the data utilization value and reconstruct the business processes including equipment inspection and safety supervision.
The fully autonomous and controllable universal intelligent computing server utilizes Limou PCIe card with a maximum computing power of 1000TOPS for a single server, supports 64-channel high-definition real-time video analysis and 10000 + FPS picture analysis, and achieves “local collection and local computing” of video information in industrial applications including smart power. It also supports multi-mode mixed running and multi-functions, and achieves real-time monitoring and early warning of equipment status, environmental risks and safety behaviors.
The pan-robot autonomous navigation and intelligent inspection solution developed based on Liyu domain controller and automatic driving software stack supports migration to inspection robots, UAV inspection and other fields. Its hardware computing power can reach 60TOPS, and the power consumption is less than 15W. It eliminates the requirement to transmit video data to the cloud, thus achieving “integration of inspection and computing” and enhancing the timeliness of inspection and reducing the high traffic fee generated by network transmission. This solution achieves real-time judgment of equipment status and potential external risks in industrial applications, and identifies abnormal situations in real-time, with rapid location and resolution.
