What is the difference between object-oriented remote sensing interpretation and traditional remote sensing image information?

Traditional remote sensing change monitoring and information extraction are mainly based on medium and low-resolution remote sensing satellite data or aerial photos. Through visual interpretation or pixel-based computer classification methods, the accuracy and efficiency of information extraction cannot be balanced.

Object-oriented remote sensing high spatial resolution images have richer spatial information, and the detailed information of ground objects can be expressed more clearly.

From a classification technology perspective

Due to the restriction of spatial resolution, traditional remote sensing image information extraction can only rely on the spectral information of the image, and it is classified at the pixel level;

Although the structure, texture and other information of object-oriented remote sensing high spatial resolution images are very prominent, the spectral information is insufficient (fewer bands). Therefore, only relying on the spectral information of pixels for classification, focusing on local pixels and ignoring the texture, structure and other information of the adjacent entire image will inevitably lead to a reduction in classification accuracy.