Weekly Wiki

Greetings

This is Frank in Seeed Studio and welcome to the Seeed Studio weekly wiki! Each Monday, I will introduce you our latest wiki, including the ones related to our new products, the interesting projects published by Seeed Studio, and the tutorials, introductions, etc... You name it.

✨ For each week, when a collaborator contributes a project or fixes something important, we put the stars at the end of title of "Weekly Wiki" for more people be able to see their efforts👍.

Today is July 29th and a brand new week has begun! Check out what Seeed Studio did last week!

Latest Wiki Launched

SenseCAP Watcher as a Grove sensor

In this wiki, we will explore the exciting possibilities that arise when Watcher, acting as a Grove sensor, leverages its UART (Universal Asynchronous Receiver/Transmitter) functionality. By enabling UART communication, Watcher can transmit valuable data such as captured images and recognition results through its UART interface, located on the back of the device. This opens up a world of opportunities for integrating Watcher with various hardware platforms and creating innovative applications.

Watcher & Node-RED To Kafka Quick Start

Apache Kafka is a distributed event streaming platform designed for high-throughput, fault-tolerant data processing. It enables real-time data feeds by allowing producers to publish messages to topics, while consumers can subscribe to these topics to process the data. Kafka is widely used for building data pipelines, real-time analytics, and integrating various data sources. Its robust architecture ensures scalability and durability, making it a popular choice for modern data-driven applications.

Why use Docker? Because Docker can simulate the environment of multiple computers on a single machine and deploy applications with great ease. Therefore, in this project, we will use Docker to set up the environment and improve efficiency.

T1000-E Tracker for Meshtastic

pir

Get One Now 🖱️

It is a high-performance tracker designed for Meshtastic, as small as a credit card, effortlessly fitting in your pocket or attaching to your assets. It embeds Semtech's LR1110, Nordic's nRF52840, and Mediatek's AG3335 GPS module, providing Meshtastic users with a high-precision, low-power positioning and communication solution.

This wiki provides a tutorial on how to use NVStreamer for RTSP video streaming on reComputer J4012.

How to Run VLM on reComputer with Jetson Platform Services

VLMs are multi modal models supporting images, video and text and using a combination of large language models and vision transformers. Based on this capability, they are able to support text prompts to query videos and images thereby enabling capabilities such as chatting with the video, and defining natural language based alerts. The VLM AI service, enables quick deployment of VLMs with Jetson Platform Services for video insight applications. The VLM service exposes REST API endpoints to configure the video stream input, set alerts and ask questions in natural language about the input video stream.

This wiki provides a tutorial on how to run VLM on reComputer J4012.

How to Run Zero-Shot Detection on reComputer with Jetson Platform Services

Generative AI vision transformers such as CLIP have made it possible to build zero shot detection models capable of open vocabulary object detection. Meaning, the model is not bounded by a set of pre-defined classes to detect. The objects to detect are configured at runtime by the user. The Zero Shot Detection AI service, enables quick deployment of generative AI with Jetson Services for open vocabulary detection on video live stream input. The Zero Shot Detection AI service exposes REST API endpoints to control stream input and objects to detect.

This wiki provides a tutorial on how to run Zero-Shot Detection on reComputer J4012.

Pose-Based Light Control with Node-Red and Raspberry Pi with AIkit

pir

This wiki will guide you on how to run YOLOv8 using an AI kit, use YOLOv8 to monitor your posture, and ultimately control your lights based on your posture. In this project, a USB camera captures your pose, and yolov8n run on AI kit with reComputer R1000 to detect your pose. The processed video, displaying the detected pose, is then streamed in real-time to reTerminal DM using gstreamer. Meanwhile, the joint coordinates are sent using mqtt to Node-RED which deploy on the reComputer R1000. At last, the Node-RED flow controls the smart lights based on the joint coordinates.

reComputer R1000 with FIN to create a Equip Graphic

FIN Framework (FIN) is a software framework with application suites that can integrate, control, manage, analyze, visualize and connect. Its capabilities can be integrated by OEMs into a range of products and services.

This article will show you how to use the Graphics Builder of FIN Framwork, and to create a Equip Graphic using the Graphics Builder.

Convert Model to Edge TPU TFlite Format for Google Coral

The Coral M.2 Accelerator with Dual Edge TPU is an M.2 module that brings two Edge TPU coprocessors to existing systems and products with an available M.2 E-key slot.Tensorflow and Pytorch is the most popular deep learning frameworks. So in order to use the Edge TPU, we need to compile the model to Edge TPU format.

This wiki article will guide you through the process of compiling a model and running it on the Google Coral TPU, enabling you to leverage its capabilities for high-performance machine learning applications.

How to Use Multiple CSI Cameras on reComputer with ROS

This tutorial provides a step-by-step guide on how to read multiple CSI camera image data through ROS on reComputer J30/J40 series devices and publish image topics to be displayed in the RVIZ visualization interface.

Recording and playback for Respeaker Lite

This project demonstrates a basic loopback mechanism using the I2S interface, to test the reading and writing functions of I2S audio data. By switching I2S mode, the audio data is read from the microphone and then written to the speaker.

How to Use NVStreamer for RTSP Streaming on reComputer with Jetson Platform Services

NVStreamer is a software developed by NVIDIA that can store and serve video files, which can then be streamed using the RTSP protocol. It is particularly useful for testing with VST, as it provides an alternative method to cameras for creating video sources as input for VST. Specifically, it offers a mechanism to use specific video files as input for VST testing. VST can be configured to receive RTSP streams as if they were coming from ONVIF-S cameras.