Object-Based Image Analysis

Object-based image analysis (OBIA) has become increasingly popular in the remote sensing field in recent years, as it allows for the creation of more meaningful and accurate classifications of land cover. Unlike traditional pixel-based methods, which classify individual pixels based solely on their spectral characteristics, OBIA groups pixels together based on their spectral, spatial, and contextual properties. By analyzing these characteristics, OBIA creates objects or features that represent real-world objects, such as buildings, vegetation, and roads. In this article, I will explore the various applications of OBIA and highlight its benefits and advantages, which have been drawn from three sources of reference. These sources demonstrate that OBIA is particularly useful in identifying complex features and objects that are difficult to distinguish using traditional pixel-based methods and that it has the potential to improve the accuracy and consistency of land cover classifications.

Source 1

The first source is a review article by Blaschke et al. (2014), which provides an overview of OBIA and its various applications. According to the authors, OBIA has several advantages over pixel-based methods, such as improved accuracy, flexibility, and interpretability. They highlight the importance of selecting appropriate segmentation and classification algorithms based on the specific application and data characteristics. The paper also includes a study that aimed to investigate the claim that GEOBIA is an important trend or paradigm in remote sensing. It involved a literature search using three different portals and found that the number of articles relevant to GEOBIA is increasing rapidly. However, it also notes that most articles discuss GEOBIA as a new paradigm without providing evidence to support their claims. The following image from the study shows the use of OBIA in different disciplines of study.

Source 2

The second source is a research article by Gao et al. (2016), which demonstrates a rule-based approach for extracting impervious surface features from high spatial resolution imagery such as IKONOS, ALOS, and SPOT-5. The authors acknowledge the challenge of extracting impervious surfaces from remotely sensed imagery with high spatial resolution but relatively few spectral bands. They propose a rule-based algorithm that can effectively extract impervious surfaces from such data and make the identification of impervious surfaces from the imagery more machine-friendly. It was interesting to note that the addition of two index bands, NDVI and NDWI, to the original four spectral bands increased the spectral dimensionality of the imagery, providing more spectral feature spaces for performing rule analysis. Also, the authors fine-tuned the algorithm’s threshold values, which helped them overcome some spectral confusion occurring in the extracted results.

Source 3

The third source is a study by Zhang et al. (2019), which used OBIA to map forest types in a complex mountainous area. In their study, Zhang et al. (2019) demonstrated the effectiveness of OBIA in mapping forest types in a complex mountainous area using OBIC, an object-based image classification method. By combining object-based segmentation with a decision tree classifier, they were able to achieve higher classification accuracy compared to traditional pixel-based methods. However, the authors also noted that the success of OBIA largely depends on the selection of appropriate features and thresholds during the segmentation process. They recommended careful consideration of these parameters to improve the accuracy of classification results. This finding highlights the importance of expert knowledge and experience in the application of OBIA, which can be used to further refine and improve the technique.

Conclusion

In summary, Object-based image analysis (OBIA) is a robust methodology that has demonstrated its effectiveness in several remote sensing domains, such as urban mapping, forest inventory, and land use/land cover mapping. The efficacy of OBIA relies on the appropriate selection of segmentation and classification techniques, inclusion of spatial and contextual information, and careful choice of features and thresholds. Additionally, incorporating expert knowledge and integrating data from multiple sources and timeframes can further enhance the accuracy of OBIA. Overall, OBIA has the capacity to advance our comprehension of the earth’s surface and provide valuable insights for decision-making in various disciplines.

References:

Blaschke, T., et al. (2014). Object-based image analysis: Is it “really” a paradigm shift? Remote Sensing, 6(4), 3636-3663.

Gao, B., et al. (2016). Object-based impervious surface mapping with high-resolution multispectral imagery. Remote Sensing, 8(5), 401.

Zhang, W., et al. (2019). A comparison of pixel-based and object-based approaches for mapping forest types in a complex mountainous area. Remote Sensing, 11(6), 715.