Stanford Team Uses AI to Locate Nearly All Solar Panels US


 
  Newest Topics Replies

 ° Ravichandran ashwin profile & biography
264
 ° Yuvraj singh profile & biography
144
 ° Munaf patel profile & biography
129
 ° Daily highlights of common wealth games cwg 2010
638
 ° India vs sri lanka match results
182
 ° Sing a song according to your own mood
310
 ° Bait bazi shairy
377
 ° Sl vs ind live score board
183
 ° Suresh raina profile & biography
98
 ° Harbhajan singh profile & biography
113
 ° Zaheer khan profile & biography
185
 ° Virender sehwag profile & biography
125
 ° Lets play antakshri
249
 ° 4 oct 2010 daily schedule updates common wealth games 2010
85
 ° Indian forthcoming movies
541
 ° Sachin tendulkar profile & biography
155
 ° Watch ipl live streaming
284
 ° Yusuf pathan profile & biography
112
 ° Gautam gambhir profile & biography
100
 ° Ms dhoni profile & biography
172
 ° Piyush chawla profile & biography
224
 ° Praveen kumar profile & biography
148
 ° Fill the blank^_^
256
 ° Say good (morning, night, afternoon)
247
 ° Indian movies release dates
203
 ° Virat kohli profile & biography
58
 ° Groxpreenok uepcm
0
 ° Astereerskef vrwfj
0
 


General discussion and chit chat
Share Bookmark and Share

Stanford Team Uses AI to Locate Nearly All Solar Panels US

Postby errohitbhardwaj on Thu Jan 23, 2020 11:34 am

To manage the changing U.S. electricity system and to understand the barriers to greater use of clean energy, it is crucial to know who have installed residential solar panels on their roofs and why.

Stanford University scientists have analysed more than a billion high-resolution satellite images using a machine learning algorithm. The algorithm has identified almost every residential solar power installation in the connecting 48 states.

The Analysis

According to the analysis, total 1.47 million installations were found, which is quite a higher number than the estimate. Scientists then combined the U.S. Census and other data with their solar catalogue so as to recognize the factors that have led to solar power adoption.

The recent advances in machine learning is facilitating scientists to know where all these assets are and generate insights about where the grid is going. The associate professor of civil and environmental engineering, Ram Rajagopal, along with professor of mechanical engineering, Arun Majumdar have been supervising the project at Stanford University.

Who Goes Solar?

The researchers found that household income played an important role. For people earning above $150,000 a year, income was playing a major role in their decisions. Low and medium income households have been found, not showing much interest in getting a residential solar system installed, despite of knowing the fact that living in areas where solar panel installation can be profitable in the long term.

For instance, areas with a relatively high electricity rates and lot of sunshine, utility bill savings would exceed the monthly cost of the equipment. According to the findings, solar installers may develop new financial models to satisfy the unmet demand.

DeepSolar – The Machine Learning Program

The team trained the machine learning program, DeepSolar, to recognize residential solar panels by providing nearly 370,000 images. Every image was labelled as either having or not having a solar panel.

Now, DeepSolar can identify features associated with solar panels, like texture, color and size. It could correctly identify an image as containing solar panels 93% of the time and missed about 10% images that had solar installations. With some novel efficiencies, the machine learning program got the job done in a month.
errohitbhardwaj
 
Posts: 25
Joined: Mon Jul 08, 2019 10:16 am

Return to Chit Chat

Who is online

Users browsing this forum: No registered users and 2 guests

cron