Rhonda creates innovative camera designs with the best camera processor (SoC) technologies, image sensor and other hardware components
Rhonda Software is a close Ambarella's design partner with many years of experience in the Imaging business, providing camera solutions based on A12S, A12W, A9SE, H2, H22, S2L, S3L, S5L, CV2, CV22
The Rhonda System on a Module (SoM) platform can accelerate your camera project or prototype. SoM is a PCB-level camera module with state-of-the-art Rhonda Image Quality and a pre-defined feature set
Rhonda is a one-stop shop for camera schematics and high-frequency HDI PCB layouts with interdisciplinary coordination with ME, optics and thermal design
Breathing life into bare metal
Highly customizable, continuously evolving Cross-platform middleware accelerates camera software development
Synchronized video capture and balanced image quality auto-correction on multi-sensor / multi-SoC cameras
Real-time video stitching for dual sensor cameras with back-to-back lens alignment
10+ Years of deep learning neural network porting and acceleration to execute complex computer vision features on-camera
Rhonda's Video Cloud solution: 2-way camera connection, archived storage, event notifications, video transcoding & streaming
Android and iOS applications for camera setup, live video streaming and over-the-air firmware updates
Ambarella CVFlow power efficient & low latency DNN inference engine and our services of making Neural Nets run on cameras
The Rhonda Software Computer Vision Team is capable of training new Neural Networks from scratch, and combining classical CV methods (such as tracking) with multiple detection and recognition DNNs into the whole system to solve customers’ unique needs | If an Ambarella-optimized DNN library is supplied by a CV partner, we can provide integration services: configure the AI library to meet quality/speed goals, enable input video at a specific resolution and FPS, implement DNN results post-processing, and delivery to external systems and decision-making |
When it comes to porting customers’ or open source AI onto the Ambarella CVflow, Rhonda’s CV Team applies its platform-specific knowledge to convert Pytorch/Tensorflow CNN/RNN into hardware accelerated representation: replacing the heavy backbone with a mobile-friendly one, substituting unsupported operations, splitting big networks into multiple modules running on the CVflow accelerator or ARM processor |
As an integral part of each CV project, On-chip validation environment and methodology allows Rhonda Software to measure accelerated DNN performance during the different stages of a project cycle | To improve DNN inference performance on the edge, Rhonda Software is ready to apply its extensive experience in DNN architecture optimization, pruning and quantization-aware re-training |
Celesta – a ready-to-deploy camera-connected Video Cloud for centralized video archiving and data storage
Deployment available both on bare metal servers and on Google Kubernetes Engine (GKE) clusters
Video records and Camera events with JPEG thumbnails. Metadata from Camera-connected sensors
Live stream from Camera or Video on Demand (VOD) from Storage: over HTTP in MPEG-DASH, RTMP and RTSP using Android / iOS SDK, MJPEG on any platform. Low latency transcoding of VOD & Live stream
REST API (JSON) with 200+ HTTP REST API resources: Multi-tenant business model, External Billing support over REST, JSON; WebSocket and Webhooks for notifications and subscriptions
Cross platform Embedded Camera Module: Windows / Linux / Embedded Linux with Real-time communication protocol using WebSocket Secure transport
A picture is worth a thousand words... What about a video?
Build an optical system most appropriate for Your application
Image sensor & lens selection for your product needs
Image Quality (IQ) Lab for thorough Optical system analysis
IQ experts configure the Ambarella SoC IQ Pipeline to deliver the best Camera IQ possible. Fully customizable IQ tuning for Computer vision applications
Rhonda's 3A (AF, AE, AWB) algorithms, Noise reduction, Tone mapping, Color fidelity adjustment, Aberrations and shading compensation, Distortion correction, Image sharpening, Local contrast enhancement, and more
Make Your camera factory process reliable and manageable
Automated testing with quantitative pass/fail criteria to find and filter out malfunctioning components at different assembly stages
Dust contamination checks, lens alignment verification, focus uniformity verification
Defective pixel correction, white balancing, lens shading compensation, etc.
Objective and consistent decision-making procedures unbiased by human-factor
Just a few highlights to tell our story
We have close relationships with the most famous names in the camera manufacturing world
Let us know how we can help you