"StarNet" High density GPU video analysis server
VT-ST40-2U is the industry's first video server based on GPU parallel structured accelerated computing, it have 40 high performance processors, and each processor board connected with high speed connector , each GPU module can be swapping, it computational capacity up to 12 TFLOPS. per processor using 28nm process design, frequency up to 2.3GHz, with 196 high-speed computing core in single processor computing capacity of up to 300 GFLOPS, while 16G EMMC has independent storage space, support multi intelligent analysis system through the formation of clusters cascade. Product support 80 1080P real-time video intelligent analysis, support vehicle behavior detection, vehicle attributes of structured information, intelligent analysis of pedestrian attribute information content; support algorithm to create module, process customization algorithm, algorithm with automatic test results and attribute module alignment.
embedded Linux operating system;
computing speed up to 12 TFLOPS;
Resolution support 1280×720（25fps）、1920×1080（25fps）、2048×1536（25fps）、2560×1920（25fps）、2560×2048（25fps）、3072×2048（25fps）ect..；
support ONVIF, GB/T28181 standard video access protocol;
support 40 independent high performance processors, support GPU parallel computing;
support 40 LED lights display the working state of each processor;
support up to 80 ways 1080P real-time video or 80 video files intelligent analysis;
supports multiple video server cluster system structured by cascading style;
support video streaming, streaming access system for intelligent analysis;
algorithm creates support module, upload module, flow separation, algorithm customization, operation flow visualization algorithm;
support algorithm with automatic test and alignment results attribute module;
support Analysis of vehicle behavior (such as: run through red light, cover line, retrograde, driver do not follow the guide lane and other violations);
support vehicle information identification (such as license plate recognition, license plate color recognition of vehicle body color, vehicle brand, vehicle type, driver's facial feature, safety belt detection etc.);
Pedestrian identity attributes recognition (such as: gender, age identification, population density monitoring etc.).