Object-oriented Image Analysis Research Articles

Waterlogged Area Identification Models Based on Object-Oriented Image Analysis and Deep Learning Methods in Sloping Croplands of Northeast China

Drainage difficulties in the waterlogged areas of sloping cropland not only impede crop development but also facilitate the formation of erosion gullies, resulting in significant soil and water loss. Investigating the distribution of these waterlogged areas is crucial for comprehending the erosion patterns of sloping cropland and preserving black soil resource. In this study, we built varied models based on two stages (one using only deep learning methods and the other combining object-based image analysis (OBIA) with deep learning methods) to identify waterlogged areas using high-resolution remote sensing data. The results showed that the five deep learning models using original remote sensing imagery achieved precision rates varying from 54.6% to 60.9%. Among these models, the DeepLabV3+-Xception model achieved the highest accuracy, as indicated by an F1-score of 53.4%. The identified imagery demonstrated a significant distinction in the two categories of waterlogged areas: sloping cropland erosion zones and erosion risk areas. The former had obvious borders and fewer misclassifications, exceeding the latter in terms of identification accuracy. Furthermore, the accuracy of the deep learning models was significantly improved when combined with object-oriented image analysis. The DeepLabV3+-MobileNetV2 model achieved the maximum accuracy, with an F1-score of 59%, which was 6% higher than that of the model using only original imagery. Moreover, this advancement mitigated issues related to boundary blurriness and image noise in the identification process. These results will provide scientific assistance in managing and reducing the impact in these places.

Potential of Lightweight Drones and Object-Oriented Image Segmentation in Forest Plantation Assessment

Forests play a vital role in maintaining ecological balance and provide numerous benefits. The monitoring and managing of large-scale forest plantations can be challenging and expensive. In recent years, advancements in remote sensing technologies, such as lightweight drones and object-oriented image analysis, have opened up new possibilities for efficient and accurate forest plantation monitoring. This study aimed to explore the utility of lightweight drones as a cost-effective and accurate method for mapping plantation characteristics in two 50 ha forest plots in the Nayla Range, Jaipur. By combining aerial photographs collected by the drone with photogrammetry and limited ground survey data, as well as topography and edaphic variables, this study examined the relative contribution of drone-derived plantation canopy information. The results demonstrate the immense potential of lightweight drones and object-oriented image analysis in providing valuable insights for optimizing silvicultural operations and planting trees in complex forest environments.

A Functional Zoning Method in Rural Landscape Based on High-Resolution Satellite Imagery

Mapping functional zones for rural landscapes is the foundational work for rural land use planning and plays a very important role in the economic development and resource management utilization of rural areas. However, the traditional manual delineation of functional zone boundaries empirically in rural areas is labor-intensive, time-consuming, and lacks the consideration of spatial landscape patterns. The emergence of high-resolution remote sensing imagery and image segmentation has facilitated the analysis of ground landscape information and patterns, but there is still a lack of functional zone boundary mapping methods applicable to rural landscapes. To address this, we propose a functional zoning method called multiscale merging of landscape contextual and shape characteristics with heterogeneity indices (M2LHI) for mapping geographic boundaries for rural landscapes based on high-resolution remote sensing imagery. The landscape contextual features were first constructed based on the geospatial distances of landscape types, and then, the dominance index and shape index were introduced to quantify the landscape heterogeneity by object-oriented image analysis. Then, the automated merging of adjacent landscape units based on the thresholds of the landscape heterogeneity indices was performed to map the initial zones. The final rural functional zones were defined based on the main function in the zone. The study was carried out in three typical rural landscapes (hilly countryside, flat countryside, and grassland countryside) located in Fujian, Xinjiang, and Inner Mongolia, China, and the freely available Gaofen-2 (GF-2) satellite imagery was used as the data source. We compared the boundaries of mapped functional zones and reference functional zones, and the matching and inclusion ratios of the final functional zones mapped in each case were bigger than 78%, indicating that the M2LHI method has a high ability to map the functional spatial patterns. The overall accuracies of mapping functional zones with different functions were 95.9%, 89.0%, and 92.1% for the respective cases. The results demonstrated that the M2LHI method effectively quantifies landscape heterogeneity and accurately delineates functional zones with different landscape patterns. It can provide a scientific basis for rural planning and management and efficiently draw reasonable geographic boundaries for rural functional zones.

Unidade de conservação no semiárido mineiro e seu papel na preservação ambiental

The monitoring of land use and occupation plays a key role in the management of Conservation Units. In this perspective, this paper presents a temporal analysis of land use and land cover for the years 1987, 1997, 2007, 2017 and 2022 in the area of the Caminho dos Gerais State Park (Parque Estadual Caminho dos Gerais – PECG) through geographic object-oriented image analysis (GEOBIA) using Landsat-TM/OLI images. The PECG is located in an ecotonal region in the semiarid north of Minas Gerais, between the municipalities of Mamonas, Monte Azul, Gameleiras, and Espinosa. In its history of occupation, this Park has presented territorial dynamics of different interests, including the impact of eucalyptus forests, until 2007, the year of its creation. The results show the effect of creating the PECG for the gradual recovery of vegetation cover in a water recharge area.

A Lake Extraction Method Combining the Object-Oriented Method with Boundary Recognition

The China–Pakistan Economic Corridor is the pilot area of the Belt and Road, where glaciers and lakes are widely distributed. Recent years, global warming has accelerated the expansion of glacier lakes, which increased the risk of natural disasters such as glacier lake outburst. It is important to monitor the glacier lakes in this region. In this paper, we propose a method combining the object-oriented image analysis with boundary recognition (OOBR) to extract lakes in several study areas of China–Pakistan Economic Corridor (CPEC). This method recognized the lake boundary with the symmetrical characteristic according to the principle of seed growth of watershed algorithm, which can correct the boundary extracted by the object-oriented method. The overall accuracy of the proposed method is up to 98.5% with Landsat series images. The experiments also show that the overall accuracy of our method is always higher than that of the object-oriented method with different segmentation scales mentioned in this paper. The proposed method improved the overall accuracy on the basis of the results obtained by the object-oriented method, and the results with the proposed method are more robust to the seeds than that with the boundary correction method of the watershed algorithm. Therefore, the proposed method can obtain a high extraction accuracy while reducing the complexity of the object-oriented extraction.

Mapping Pu'er tea plantations from GF-1 images using Object-Oriented Image Analysis (OOIA) and Support Vector Machine (SVM).

Tea is the most popular drink worldwide, and China is the largest producer of tea. Therefore, tea is an important commercial crop in China, playing a significant role in domestic and foreign markets. It is necessary to make accurate and timely maps of the distribution of tea plantation areas for plantation management and decision making. In the present study, we propose a novel mapping method to map tea plantation. The town of Menghai in the Xishuangbanna Dai Autonomous Prefecture, Yunnan Province, China, was chosen as the study area, andgg GF-1 remotely sensed data from 2014-2017 were chosen as the data source. Image texture, spectral and geometrical features were integrated, while feature space was built by SEparability and THresholds algorithms (SEaTH) with decorrelation. Object-Oriented Image Analysis (OOIA) with a Support Vector Machine (SVM) algorithm was utilized to map tea plantation areas. The overall accuracy and Kappa coefficient ofh the proposed method were 93.14% and 0.81, respectively, 3.61% and 0.05, 6.99% and 0.14, 6.44% and 0.16 better than the results of CART method, Maximum likelihood method and CNN based method. The tea plantation area increased by 4,095.36 acre from 2014 to 2017, while the fastest-growing period is 2015 to 2016.

Analysis on Dynamic Changes of Roliai Laterite Nickel Mine from 2013 to 2020

In order to accurately obtain the mining activities of Roliai laterite nickel mine, based on the remote sensing technology and object-oriented image analysis, methods of multi-scale segmentation based on scale optimization and optimal classification threshold selection were used to extract nine mining activity elements from WorldView II images of of 2013 and 2020 respectively. According to the extraction results, the dynamic changes of each element from 2013 to 2020 were analysed from the aspects of area and intensity. Our accuracy evaluation results showed that the overall accuracies of the two extraction results are higher than 90%, and the Kappa coefficients are higher than 0.81. From 2013 to 2020, the mining scale and mining intensity of Roliai laterite nickel mine showed a trend of rapid expansion on the whole, in which the mining area increased the most, reaching 1.02 km2, and the average annual change intensity was intense, and the corresponding dump area also increased by about 0.35 km2. In the past seven years, Roliai laterite nickel mine has developed from the stage of exploration and mining to forming a relatively complete industrial chain of mining, milling and metallurgyy, entering a development stage of rapid expansion.

Mapping Irregular Local Climate Zones from Sentinel-2 Images Using Deep Learning with Sequential Virtual Scenes

Recently, the local climate zone (LCZ) system has been presented to establish the connection between urban landscape and local thermal environment. However, LCZ entities are very difficult to be identified by pixel-based classifiers or object-oriented image analysis, as they are often a complicated combination of multiple ground objects (e.g., buildings, roads, grassland, etc.). Scene classifiers, especially deep learning methods can exploit the structure or contextual information of image scenes and then improve the performance of LCZ classification. However, the square and uniform-sized image patches often bring about extra challenges, as they cannot exactly match LCZ entities of diverse sizes and shapes in most cases. In this study, a sequential virtual scene method is presented to identify LCZ entities of diverse shapes and sizes, which consists of a small “core patch” for scanning diverse entities and sequential virtual scenes for providing abundant context. Specifically, the Bidirectional Long Short-Term Memory (Bi-LSTM) were used to learn the spatial relationship among virtual scenes, respectively. Importantly, a “self-attention” mechanism is designed to weigh the contribution of every virtual scene for alleviating the influences of mixed patches, according to the similarity between its hidden state and the final hidden state. Experiments prove SVS achieves better accuracies than random forest and ResNet and has the outstanding capacity of identifying irregular LCZ entities. It is a promising way to carry out LCZ mapping in cities of different types due to its flexibility and adaptability.

3D MAPPING OF BURIED PIPES IN MULTI-CHANNEL GPR DATA

Abstract. Ground Penetrating Radar (GPR) allows a non-destructive analysis of the subsurface by using electromagnetic waves. GPR is used in application fields such as archeology and civil engineering, where the detection of buried objects is in demand. Such objects, e.g. pipes, lead to a disturbance of the propagation of the radar signal in the underground. A manual detection of disturbing objects can be both time-consuming and tedious, depending on the number of GPR images to be investigated. Hence, automatic methods should be considered. In this study, an object-oriented image analysis approach for the automatic detection and mapping of buried pipes is presented and evaluated. As dataset, measurements of the multi-channel system Stream C are considered.

Long-Term Monitoring of Vegetation Dynamics in the Rhodopi Mountain Range National Park-Greece

While deforestation is a major environmental issue in the tropics, with thousands of hectares converted to agricultural land every year, in Europe the opposite trend is observed, with land abandonment in mountainous and semi-mountainous areas allowing the afforestation of former agricultural and pastoral land. This trend allows semi-natural ecosystems to recover after a prolonged period of exploitation and often over-exploitation, but it may also lead to significant loss of landscape heterogeneity with potentially detrimental effects on biodiversity. The current study aims to monitor changes in the vegetation coverage across a period of 35 years (between 1984 and 2019) in the Rhodopi Mountains range National Park in northern Greece. A time series of LANDSAT TM (16 images), LANDSAT ETM + (1 image) and LANDSAT 8 OLI/TIRS (4 images) were employed. One data transformation method was applied (TCT), and five vegetation indices (NDVI, NDWI, SAVI, EVI2 and BSI) were calculated to capture the land cover transition during the study period. The obtained results and all used indices suggest that over the study period there was a continuous trend of vegetation cover increasing, with open areas decreasing. The observed trend was further confirmed using Object Oriented Image Analysis on two pairs of images sensed in 1984 and 2019, respectively. The results suggest that almost 22.000 ha of open habitats have been lost to broadleaved and conifer woodlands, while the former also appear to be advancing into conifer-covered areas. This trend has led to significant loss of landscape heterogeneity and to a broadleaf-dominated landscape. The results are discussed in relation to their driving forces, the potential implications on biodiversity and the risk of wildfires in the near future.

Greenhouse extraction with high-resolution remote sensing imagery using fused fully convolutional network and object-oriented image analysis

The wide application of agricultural greenhouses globally has brought economic benefits; however, it has also led to many environmental problems. The timely and accurate acquisition of greenhouse areas and distribution is valuable for authorities seeking to optimize regional agricultural management and mitigate environmental pollution. Automatic extraction of the greenhouses from high spatial resolution remote sensing (RS) imagery based on deep learning can reduce labor costs and improve operational efficiency to have better application prospects. In this paper, we propose a multi-channel fused fully convolutional network (FCN) optimized by the optimal scale object-oriented segmentation results for agricultural greenhouse extraction from high spatial resolution RS imagery. First, to make full use of remote sensing feature images of target objects and to not increase the complexity of the deep learning network, we constructed a decision-level fusion FCN network that can simultaneously input multiple remote sensing images for preliminary extraction of greenhouse. Second, to address a defect in the classical FCN network causing the easy loss of ground object details, we optimized the preliminary extraction results from FCN by the results of object-oriented segmentation. Finally, the optimized greenhouse extraction results were processed by the mathematical morphology, and the final extraction results were obtained. The experimental results demonstrate that: (1) Multi-channel fused FCN model could use the unique spectral characteristics of different ground objects. (2) Optimized initial extraction results from FCN based on the optimal scale object-oriented segmentation results could fully maintain the edge details of the greenhouse. Experimental results show that the proposed method can extract the greenhouse effectively. The precision and F value of our proposed method are 92.68% and 0.94.

Quantification of Loess Landforms from Three-Dimensional Landscape Pattern Perspective by Using DEMs

Quantitative analysis of the differences and the exploration of the evolution models of different loess landform types are greatly important to the in-depth understanding of the evolution process and mechanism of the loess landforms. In this research, several typical loess landform areas in the Chinese Loess Plateau were selected, and the object-oriented image analysis (OBIA) method was employed to identify the basic loess landform types. Three-dimensional (3D) landscape pattern indices were introduced on this foundation to measure the morphological and structural features of individual loess landform objects in more detail. Compared with the traditional two-dimensional (2D) landscape pattern indices, the indices consider the topographic features, thereby providing more vertical topographic information. Furthermore, the evolution modes between different loess landform types were discussed. Results show that the OBIA method achieved satisfying classification results with an overall accuracy of 88.12%. There are evident differences in quantitative morphological indicators among loess landform types, especially in indicators such as total length of edge, mean patch size, landscape shape index, and edge dimension index. Meanwhile, significant differences are also found in the combination of loess landform types corresponding to different landform development stages. The degree of surface erosion became increasingly significant as loess landforms developed, loess tableland area rapidly reduced or even vanished, and the dominant loess landform types changed to loess ridge and loess hill. Hence, in the reconstruction and management of the Loess Plateau, the loess tableland should be the key protected loess landform type. These preliminary results are helpful to further understand the development process of loess landforms and provide a certain reference for regional soil and water conservation.

Detection of Flood Damage in Urban Residential Areas Using Object-Oriented UAV Image Analysis Coupled with Tree-Based Classifiers

Timely clearing-up interventions are essential for effective recovery of flood-damaged housing, however, time-consuming door-to-door inspections for insurance purposes need to take place before major repairs can be done to adequately assess the losses caused by flooding. With the increased probability of flooding, there is a heightened need for rapid flood damage assessment methods. High resolution imagery captured by unmanned aerial vehicles (UAVs) offers an opportunity for accelerating the time needed for inspections, either through visual interpretation or automated image classification. In this study, object-oriented image segmentation coupled with tree-based classifiers was implemented on a 10 cm resolution RGB orthoimage, captured over the English town of Cockermouth a week after a flood triggered by storm Desmond, to automatically detect debris associated with damages predominantly to residential housing. Random forests algorithm achieved a good level of overall accuracy of 74%, with debris being correctly classified at the rate of 58%, and performing well for small debris (67%) and skips (64%). The method was successful at depicting brightly-colored debris, however, was prone to misclassifications with brightly-colored vehicles. Consequently, in the current stage, the methodology could be used to facilitate visual interpretation of UAV images. Methods to improve accuracy have been identified and discussed.

Derivation of spatially detailed lentic habitat map and inventory at a basin scale by integrating multispectral Sentinel-2 satellite imagery and USGS Digital Elevation Models

Aquatic habitats are essential for aquaculture, agriculture, and municipal water supply. The geographical location and quantitative information about the aquatic habitats are critical for inventorying, monitoring, and managing these valuable resources. This paper presents a novel method that integrates the object-oriented image analysis and DEM-based stream network analysis to map and quantify lentic aquatic habitats, using freely available ESA Sentinel-2 multispectral imagery and USGS DEMs. This integrated method has been applied to the entire Mobile River Basin, USA, and lentic habitats at the basin scale have been delineated and inventoried at the 10-m spatial resolution. The surface water pixels of aquatic habitats are extracted through the Iterative Self-Organizing Data Analysis Technique (ISODATA) unsupervised classification of the Sentinel-2 multispectral images. An object-oriented image analysis approach is then used to group contiguous water pixels to form and map discrete aquatic habitat objects. The stream network in the basin has been hydrologically derived from the USGS DEMs at the 10-m spatial resolution. The combination of the geometric and shape information from the object-oriented image analysis and hydrological information from the DEM-based stream network analysis enables the separation of lotic and lentic habitats. A set of planimetric, geometric, and shape attributes has been numerically derived for each lentic habitat object. The abundance pattern, shape pattern, and spatial distribution pattern of the lentic habitats in the basin have been analyzed. The vast majority of the aquatic habitat objects identified in the Mobile River Basin are lentic habitats. The most striking spatial pattern of lentic habitats is that small, oblong lakes and fish ponds are densely distributed along the Black Prairie Belt in the basin, laying the foundation for the catfish aquacultural industry in the regions of mid-west Alabama and east Mississippi. This research demonstrates that the integration of the recently emerged Sentinel-2A/B satellite multispectral imagery with USGS DEMs represents an effective approach to the inventory and environmental investigation of lentic habitats at a basin scale.

Improving object-oriented land use/cover classification from high resolution imagery by spectral similarity-based post-classification

To classify an image, traditional classifiers depend mainly on the spectral and/or textural distinctions between different land-cover units, while this study attempts to explore the properties of statistical distinction. Using the historical classification results, we present a novel algorithm for imagery classification that achieves high accuracy, automation and efficiency. Based on object-oriented image analysis, it exploits the advantages of d ch (the Chaudhuri’s metric) using a multi-step approach, and the objective is not to reclassify an image, but to refine or update the existing land-use/-cover classification results by comparing the pairwise d ch value (namely similarity) between different image segments. Finally, the similar/homogeneous segments will be confirmed as their original class labels, while the inhomogeneous/dissimilar segments will be masked out with an appropriate threshold on the similarity image and be relabelled. We have systematically evaluated the algorithm by running it on the basis of the existing GIS base maps, which indicated the good performance of it.

Depth Semantic Segmentation of Tobacco Planting Areas from Unmanned Aerial Vehicle Remote Sensing Images in Plateau Mountains

The tobacco in plateau mountains has the characteristics of fragmented planting, uneven growth, and mixed/interplanting of crops. It is difficult to extract effective features using an object-oriented image analysis method to accurately extract tobacco planting areas. To this end, the advantage of deep learning features self-learning is relied on in this paper. An accurate extraction method of tobacco planting areas based on a deep semantic segmentation model from the unmanned aerial vehicle (UAV) remote sensing images in plateau mountains is proposed in this paper. Firstly, the tobacco semantic segmentation dataset is established using Labelme. Four deep semantic segmentation models of DeeplabV3+, PSPNet, SegNet, and U-Net are used to train the sample data in the dataset. Among them, in order to reduce the model training time, the MobileNet series of lightweight networks are used to replace the original backbone networks of the four network models. Finally, the predictive images are semantically segmented by trained networks, and the mean Intersection over Union (mIoU) is used to evaluate the accuracy. The experimental results show that, using DeeplabV3+, PSPNet, SegNet, and U-Net to perform semantic segmentation on 71 scene prediction images, the mIoU obtained is 0.9436, 0.9118, 0.9392, and 0.9473, respectively, and the accuracy of semantic segmentation is high. The feasibility of the deep semantic segmentation method for extracting tobacco planting surface from UAV remote sensing images has been verified, and the research method can provide a reference for subsequent automatic extraction of tobacco planting areas.

Spatial differentiation characteristics and driving mechanism of rural-industrial Land transition: A case study of Beijing-Tianjin-Hebei region, China

Given China's economic development, its rural areas have experienced rapid and profound transitions. In the Beijing-Tianjin-Hebei (BTH) region, these transitions have resulted in transition in rural-industrial land. Therefore, with rural-industrial land transition being the driving force behind rural economic growth, discerning its characteristics and driving mechanism could provide insights needed for guiding its regulation as well as determining suitable rural development paths. After constructing theoretical models for the rural-industrial land temporal and spatial transition in the BTH region, this study empirically analyzes the spatial differentiation characteristics and driving mechanism of rural-industrial land transition using object-oriented image analysis, the comprehensive index method, ArcGIS and FRAGSTATS landscape pattern analysis, and ordinary least squares and geographical weighted regression model. The results indicate that the high values of rural-industrial land dominant and recessive transition are distributed in the economically developed regions—Beijing-Tianjin and Handan—while undeveloped regions, such as Xingtai and Hengshui regions, show low values. Moreover, the high-value areas of the rural-industrial land transition index (RILT) are striped and the low-value areas are clustered. Furthermore, the spatial heterogeneity of rural-industrial land transition patterns is apparent. The transition coordination regions are located in the Beijing-Tianjin regions, the recessive lagging patterns are concentrated in the Langfang and Cangzhou regions, the dominant transition lagging regions are located in the Handan regions, and the transition lagging regions distributed in Xingtai and Hengshui regions. Finally, the results reveal that rural-industrial land transition is primarily driven by population, socio-economic, and location factors. For example, the degree of a rural-industrial land transition is positively related to socio-economic growth level. Thus, properly coordinating the relationship between these two is conducive to coordinating the BTH region’s development, promoting regional rural revitalization and sustainable rural development in China. To facilitate this, targeted transition paths for different transition regions are therefore proposed.

Object-based approach for urban land cover mapping using high spatial resolution data

This paper deals with object-oriented image analysis applied for an urban area. Very high-resolution images in conjunction with object-oriented image analysis have been used for land cover detection. Using the eCognition software with object-oriented methods, not only the spectral information but also the shape, compactness and other parameters can be used to extract meaningful objects. The spectral and geometric diversity of urban surfaces is a very complex research issue. It is the main reason why additional information is needed to improve the outcome of classification. The most consistent and relevant characteristic of buildings is their height. Therefore, elevation data (converted from LIDAR data) are used for building extraction, segmentation and classification. The study deals with the problem, how to determine the most appropriate parameters of segmentation, feature extraction and classification methods. The data extraction includes two phases, the first part consists the following steps: data pre-processing, rule set development, multi-scale image segmentation, the definition of features used to map land use, classification based on rule set and accuracy evaluation. The second part of the data process based on classical raster analysis GIS tools like focal and zonal function.

GEOBIA, TREE DECISION AND HIERARCHICAL CLASSIFICATION FOR MAPPING GULLY EROSION

The gullies provoke environmental, social and financial damages. The application of corrective and preventive measures needs gullies mapping and monitoring. In this scope, this study proposes a methodology for gullies delimitation using object-oriented image analysis. For such, there were used high spatial resolution imagery and ALS data applied for two study areas, one in Uberlandia-Minas Gerais (Brazil) and another one in Queensland (Australia). The objects were generated by multiresolution segmentation. The most important attributes on the delimitation of the gullies were selected using decision tree induction algorithms, being them: spectral, altimetric and texture. Classifications by decision trees and hierarchical were carried out. The use of decision tree allowed the selection of attributes and the establishment of preliminary decision rules. However, since this procedure did not use fuzzy logic, mixtures between classes could not be evidenced in the rule base. Moreover, the classification was performed by a factor of scale only, which did not allow the identification of all the constituent features of the gully. In hierarchical classification, the procedure is performed on different scales, allowing the use of fuzzy logic to describe different degrees of membership in each class, which makes it a very attractive method for cases such as this study, where there is mixing of classes. The classification obtained with hierarchical classification it was more reliable with the field truth, by allowing the use of different scales, uncertainty insert and integration of knowledge, compared to the automatic classification by decision tree.

Comparison of machine learning algorithms for mapping mango plantations based on Gaofen-1 imagery

Mango is a commercial crop on Hainan Island, China, that is cultivated to develop the tropical rural economy. The development of accurate and up-to-date maps of the spatial distribution of mango plantations is necessary for agricultural monitoring and decision management by the local government. Pixel-based and object-oriented image analysis methods for mapping mango plantations were compared using two machine learning algorithms (support vector machine (SVM) and Random Forest (RF)) based on Chinese high-resolution Gaofen-1 (GF-1) imagery in parts of Hainan Island. To assess the importance of different features on classification accuracy, a combined layer of four original bands, 32 gray-level co-occurrence (GLCM) texture indices, and 10 vegetation indices were used as input features. Then five different sets of variables (5, 10, 20, and 30 input variables and all 46 variables) were classified with the two machine learning algorithms at object-based level. Results of the feature optimization suggested that homogeneity and variance were very important variables for distinguishing mango plantations patches. The object-based classifiers could significantly improve overall accuracy between 2–7% when compared to pixel-based classifiers. When there were 5 and 10 input variables, SVM showed higher classification accuracy than RF, and when the input variables exceeded 20, RF showed better performances. After the accuracy achieved saturation points, there were only slightly classification accuracy improvements along with the numbers of feature increases for both of SVM and RF classifiers. The results indicated that GF-1 imagery can be successfully applied to mango plantation mapping in tropical regions, which would provide a useful framework for accurate tropical agriculture land management.