Change detection in landscapes - メーカー・企業と製品の一覧

Detecting changes in landscapes from in-vehicle cameras [Saitama University]

Images distributed in space and time can be effectively utilized! There is novelty and superiority in extracting the characteristics of the landscape.

We will introduce a technology developed at Saitama University for accurately detecting spatiotemporal phenomena arising from local climate and environmental conditions. By acquiring multiple imaging data sets that include coordinate and timestamp data, we calculate the probability that specific features predetermined within the imaging data are present using a deep learning model. In the real world, we calculate the average probability computed daily within a set grid and detect or estimate changes in phenomena from the resulting time series data. 【Features】 - It is possible to understand landscapes from imaging data with location information, such as from onboard cameras. - This technology enables the understanding of regional environments and landscapes that cannot be captured by satellite observations or drones. - It is expected to be useful for detecting disaster damage, abandoned farmland, deforestation, land cover changes, landscape changes, and heat environment changes such as urban heat islands. *For more details, please download the PDF or feel free to contact us.