Dahua Technology, a global leader in video-centric AI solutions for the Internet of Things (IoT), has launched its Xinghan large-scale AI models. Designed to address complex real-world challenges, the Xinghan system combines visual intelligence with multimodal and language processing capabilities, marking a significant step in Dahua’s continuous innovation and industrial AI transformation.
Core Architecture and Model Families
The Xinghan AI system is built around visual analysis and deep industry expertise, providing scalable and commercially viable AI solutions. Its architecture features three core model families:
-
L Models: Focused on natural language understanding and interaction.
-
Rad V Models (Xinghan Vision): Specialize in advanced visual intelligence and video analytics, streamlining detection of people, motor vehicles, and non-motor vehicles with high accuracy.
-
M Series: Multimodal AI models capable of processing text, images, audio, and video simultaneously, enabling natural human-computer interaction and diverse application scenarios.
Advanced Visual Intelligence Features
The Rad V model line introduces a suite of next-generation capabilities:
-
Perimeter Protection: Detects smaller targets (up to 20×20 pixels), reduces false positives, and extends camera coverage.
-
Wiz Tracking: Enhances intelligent tracking algorithms to handle occlusions and changes in posture, achieving 50% improved accuracy.
-
Crowd Map: Detects small targets at longer distances, supports up to 5,000 people, and improves performance in low-light or rainy conditions.
-
Adaptive Scenes: Uses situational awareness to intelligently adjust camera configuration.
-
AI Rule Assistance: Automates perimeter protection rule setup with high accuracy and one-click configuration.
Multimodal Intelligence for IoT
The M Series multimodal models significantly improve efficiency in processing heterogeneous data. Key features include:
-
WizSeek: Enables natural language video search across archived footage by describing targets in simple language.
-
Text-Defined Alarms: Allows users to configure alarms using natural language, facilitating flexible, scalable, and fast deployment.
Driving Intelligent Transformation
By integrating edge-cloud synergy with industry-focused AI, the Xinghan models empower intelligent, adaptive, and scalable AI across sectors. The system exemplifies Dahua Technology’s commitment to continuous development and the intelligent future of IoT.
