Cloud Sky Classifier

Introduction

Portable particulate sensors are becoming quite popular in consumer markets. They usually rely on laser scattering technology. The current document proposes a device designed to gain particulate matter measurements(pm1, pm2.5 pm10) relying on the color of the sky which may lead to unexplored atmospheric conclusions on air quality.

Design Specifications

1. Sensor Brain: Odroid XU4 -The XU4 runs on Ubuntu 16.04

2. Main Housing:

• The main cubicle enclosure will be equipped with a transparent lid.

3. Portable USB Camera:

• The 2.1mm Wide Angle Mjpeg 5 megapixel Hd Camera HD USB Cam

4. Power Usage:

• Uses between 3W - 12W

5. Date Storage:

• The device stores both the original and classified images and the classified images within an onbaord SD Card. On future iterations, both images will be uploaded to google drive via the google drive API

6 - Web Connectivity:

• The node will have a hardwired internet connection via a RJ45, Ethernet cable.

7. Date Outage:

• The device sends data packets to google drive doc which is updated live. A sample of such a document is given below.

dateTime	cloudPecentage	skyRed	skyBlue	skyGreen	cloudRed	cloudGreen	cloudBlue
2018-11-16-1-3-52	99.9156	62.5550	19.7144	9.3899	206.7500	164.4323	163.6752
2018-11-16-1-36-11	99.9082	61.7375	19.7446	8.7836	205.9728	163.5578	163.3363
2018-11-16-1-36-45	99.9010	64.6513	20.2039	8.1907	206.1724	163.5491	162.2134
2018-11-17-0-39-35	87.5651	42.4589	35.5525	9.7457	118.6922	112.5238	97.7310

8. Weight

• weighs 0.8 Kg.