CN

Solutions

Home > Solutions > Edge Computing & Robotics
Edge Computing & Robotics

Edge Computing & Robotics

These solutions are designed to enhance machine vision and drive innovation across various consumer and industrial products, such as household products, smart phones, and cameras. Applications include robotic vacuums, quick boot battery cameras, smart home AI Box, and industrial robots.

Use Cases

Night Vision Pre-ISP

Night Vision Pre-ISP

AI-ISP chips can be introduced to traditional smart phone SoCs. This technology achieves the “1 frame, 2 chips" functionality, enhancing image quality with AI-NR and AI-HDR.
Stabilized Aerial Cameras

Stabilized Aerial Cameras

Featuring a self-developed EIS algorithm, these cameras stabilize the image capturing by reducing the influence from flight attitude and propeller vibrations. This cuts down the cost, size, and weight of tripod heads on lightweight aerial devices while im
Quick Boot Battery Cameras

Quick Boot Battery Cameras

Featuring proprietary low-power standby and quick boot technology, these cameras capture an image in just 0.5 seconds, with outstanding performance in both night vision mode and black-light full-color mode, significantly extending operational time.
Smart Home AI Box

Smart Home AI Box

Capable of connecting to various video sources, this AI box enables high-frame-rate AI video detection, facilitating smart home control.
Robotic Vacuum

Robotic Vacuum

Leveraging a self-developed NPU model and a comprehensive toolchain for rapid model deployment, these robots support the integration of multiple 3D sensors and algorithm deployment while optimizing power consumption through hardened CV operators.
Robotic Lawn Mower

Robotic Lawn Mower

Capable of processing visual NN algorithms in challenging lighting conditions, these robots accurately locate lawn edges and assist in obstacle avoidance. The visual algorithms also enhance the robustness of outdoor positioning.
Smart Service Robots

Smart Service Robots

Powered by a self-developed NPU, these robots demonstrate exceptional support and efficiency for spatial perception and localization models such as BEV, NN depth, and NN SLAM, all based on CNN architecture.
Multi-modal Multi-dimensional Perception Model

Multi-modal Multi-dimensional Perception Model

Detects unknown objects and automates image labeling.
AI Image Segmentation

AI Image Segmentation

A breakthrough in computer vision. Enables image segmentation without the need for manual labeling.
Potent AI Duo

Potent AI Duo

Generates text from images, making image analysis more intelligent.
Searching for Images with Text

Searching for Images with Text

This intelligent feature supports searching for images with text.
Latest AI Feature: Monocular Depth Estimation

Latest AI Feature: Monocular Depth Estimation

Effectively estimates the depth map in a 2D image
AI Agent

AI Agent

Small chip drives human instruction large model for accurately understanding and generating human language.
32-ch Video

32-ch Video

Real-time AI detection with total power consumption of the board less than 8 W.

Universal AI for a Better World
Copyright ©2001-2022 Axera Semiconductor Co., LTD. Beijing ICP No. 19031994-4