PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
Ground object detection is important for many civilian applications. Counting the number of cars in parking lots can provide very useful information to shop owners. Tent detection and counting can help humanitarian agencies to assess and plan logistics to help refugees. In this paper, we present some preliminary results on ground object detection using high resolution Worldview images. Our approach is a simple and semi-automated approach. A user first needs to manually select some object signatures from a given image and builds a signature library. Then we use spectral angle mapper (SAM) to automatically search for objects. Finally, all the objects are counted for statistical data collection. We have applied our approach to tent detection for a refugee camp near the Syrian-Jordan border. Both multispectral Worldview images with eight bands at 2 m resolution and pansharpened images with four bands at 0.5 m resolution were used. Moreover, synthetic hyperspectral (HS) images derived from the above multispectral (MS) images were also used for object detection. Receiver operating characteristics (ROC) curves as well as detection maps were used in all of our studies.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
The alert did not successfully save. Please try again later.
Chiman Kwan, Bryan Chou, David Gribben, Leif Hagen, Jerry Yang, Bulent Ayhan, Krzysztof Koperski, "Ground object detection in worldview images," Proc. SPIE 11018, Signal Processing, Sensor/Information Fusion, and Target Recognition XXVIII, 1101817 (7 May 2019); https://doi.org/10.1117/12.2518529