Remote Sensing Image Land Classification Based on Deep Learning

Aiming at the problems of high-resolution remote sensing images with many features and low classification accuracy using a single feature description, a remote sensing image land classification model based on deep learning from the perspective of ecological resource utilization is proposed. Firstly, the remote sensing image obtained by Gaofen-1 satellite is preprocessed, including multispectral data and panchromatic data. Then, the color, texture, shape, and local features are extracted from the image data, and the feature-level image fusion method is used to associate these features to realize the fusion of remote sensing image features. Finally, the fused image features are input into the trained depth belief network (DBN) for processing, and the land type is obtained by the Softmax classifier. Based on the Keras and TensorFlow platform, the experimental analysis of the proposed model shows that it can clearly classify all land types, and the overall accuracy, F1 value, and reasoning time of the classification results are 97.86%, 87.25%, and 128 ms, respectively, which are better than other comparative models.

Full-learning quaternion convolutional neural networks and cross-entropy-based confluence of differently represented data for PolSAR land classification

ALOS2 Dataset1<div><div>Odi_SceneId="SARD000000088012-00036-019-001"</div><div>Odi_SiteDateTime="PROCESS:JAPAN-JAXA-ALOS2-EICS 20151207 072121"</div><div>Scs_SceneID="ALOS2066310850-150815"</div></div><div><br></div><div>ALOS2 Dataset2</div><div><div>Odi_SceneId="SARD000000088012-00036-044-002"</div><div>Odi_SiteDateTime="PROCESS:JAPAN-JAXA-ALOS2-EICS 20210113 054828"</div><div>Scs_SceneID="ALOS2066310860-150815"</div></div>

Full-learning quaternion convolutional neural networks and cross-entropy-based confluence of differently represented data for PolSAR land classification

ALOS2 Dataset1<div><div>Odi_SceneId="SARD000000088012-00036-019-001"</div><div>Odi_SiteDateTime="PROCESS:JAPAN-JAXA-ALOS2-EICS 20151207 072121"</div><div>Scs_SceneID="ALOS2066310850-150815"</div></div><div><br></div><div>ALOS2 Dataset2</div><div><div>Odi_SceneId="SARD000000088012-00036-044-002"</div><div>Odi_SiteDateTime="PROCESS:JAPAN-JAXA-ALOS2-EICS 20210113 054828"</div><div>Scs_SceneID="ALOS2066310860-150815"</div></div>

Decision tree-based classification in coastal area integrating polarimetric SAR and optical data

PurposeAs one of the world's most productive ecosystems, ecological land plays an important role in regional and global environments. Utilizing advanced optical and synthetic aperture radar (SAR) data for land cover/land use research becomes increasingly popular. This research aims to investigate the complementarity of fully polarimetric SAR and optical imaging for ecological land classification in the eastern coastal area of China.Design/methodology/approachFour polarimetric decomposition methods, namely, H/Alpha, Yamaguchi3, VanZyl3 and Krogager, were applied to Advanced Land Observing Satellite (ALOS) SAR image for scattering parameter extraction. These parameters were merged with ALOS optical parameters for subsequent classification using the object-based quick, unbiased, efficient statistical tree decision tree method.FindingsThe experimental results indicate that an improved classification performance was obtained in the decision level when merging the two data sources. In fact, unlike classification using only optical images, the proposed approach allowed to distinguish ecological land with similar spectrum but different scattering. Moreover, unlike classification using only polarimetric information, the integration of polarimetric and optical data allows to accurately distinguish reed from artemisia and sand from salt field and therefore achieve a detailed classification of the coastal area characteristics.Originality/valueThis research proposed an integrated classification method for coastal ecological land with polarimetric SAR and optical data. The object-based and decision-level fusion enables effective ecological land classification in coastal area was verified.

Natural disturbance regimes for implementation of ecological forestry: a review and case study from Nova Scotia, Canada

Ecological forestry is based on the idea that forest patterns and processes are more likely to persist if harvest strategies produce stand structures, return intervals, and severities similar to those from natural disturbances. Taylor et al. (2020) reviewed forest natural disturbance regimes in Nova Scotia, Canada, to support implementation of ecological forestry. In this follow-up paper, we 1) review use of natural disturbance regimes to determine target harvest rotations, age structures, and residual stand structures; and 2) describe a novel approach for use of natural disturbance regimes in ecological forestry developed for Nova Scotia. Most examples of ecological forestry consider only the local, dominant disturbance agent, such as fire in boreal regions. Our approach included: 1) using current ecological land classification to map potential natural vegetation (PNV) community types; 2) determining cumulative natural disturbance effects of all major disturbances, in our case fire, hurricanes, windstorm, and insect outbreaks for each PNV; and 3) using natural disturbance regime parameters to derive guidelines for ecological forestry for each PNV. We analyzed disturbance occurrence and return intervals based on low, moderate, and high severity classes (<30, 30-60, and >60% of biomass of living trees killed), which were used to determine mean annual disturbance rates by severity class. Return intervals were used to infer target stand age-class distributions for high, moderate, and low severity disturbances for each PNV. The range of variation in rates of high severity disturbances among PNVs was from 0.28% yr-1 in Tolerant Hardwood to 2.1% yr-1 in the Highland Fir PNV, equating to return intervals of 357 years in Tolerant Hardwood to 48 yrs in Highland Fir PNVs. As an example, this return interval for the Tolerant Hardwood PNV resulted in target rotation lengths of 200 years for 35% of the PNV area, 500 years for 40%, and 1000 years for 25%. The proposed approach of determining natural disturbance regimes for PNV communities and calculating target disturbance rates and corresponding harvest rotation lengths or entry times appears to be a feasible method to guide ecological forestry in any region with a strong ecological land classification system and multiple disturbance agents.

Geotechnology applied to land classification in areas of the Atlantic Forest in southeastern Brazil

The objective of this study was to classify the lands of a micro-watershed located in the Atlantic forest biome, in a region of rough relief, in the use capacity system using geotechnology resources and indicate uses for the lands according to their suitability. The theoretical basis of the Manual for Utilitarian Survey and Classification of Land in the Use Capacity System with adaptations for areas of rough relief was adopted. The study was carried out from the survey of topographic information to construct the altimetric map of the watershed, followed by the survey of the physical environment, especially water erosion, description of soil profiles and collection of samples. The parameters effective depth, texture, permeability, slope, erosion, fertility, and land use were evaluated. Based on the pedological data and on the use of applied geotechnology, the soil map was created, and the lands of the watershed were classified and mapped in the use capacity system. After interpretation of the survey products, it was verified that in rough relief, slope is the predominant factor to determine the classes of land use, as it outweighs the other parameters evaluated. Land classification land with the use capacity system promotes optimization in the use of areas with agricultural areas and preservation of those destined for conservation.