Change Detection Algorithm Research Articles

Automated Forest Harvest Detection With a Normalized PlanetScope Imagery Time Series

The advent of CubeSat constellations is revolutionizing the ability to observe Earth systems through time. The improved spatial and temporal resolutions from these data could assist in tracking forest harvesting by forest management companies or government organizations interested in monitoring the sustainable management of forest resources. However, differing characteristics of individual satellites in each constellation requires study into geometric and radiometric normalization of the imagery and tuning parameters for change detection algorithms. In this study, a method for the spatial and temporal detection of forest harvest operations using images from the PlanetScope constellation was developed and implemented for a managed forest in Ontario, Canada. Temporal smoothing was applied on Landsat-normalized PlanetScope values of the Normalized Differential Vegetation Index (NDVI), and change points were detected based on the first derivative of the NDVI trend. Detected changes were compared to known locations of harvesting machines. Results indicate that 80–90% of harvested areas were detected, with temporal errors of approximately 9–10 days for two sites. Overall, this study demonstrated that forest harvesting can be detected with relative accuracy, deriving previously unavailable levels of spatial and temporal detail and enhancing the ability of forest stakeholders to monitor the sustainable use of forest resources.

Testing the detection of large, secretive snakes using camera traps

AbstractNovel technologies, such as camera traps, have expanded the opportunities for species detection, especially for rare species. Corresponding changes in data processing must occur to handle the large volume of data gathered from technology like camera traps. Automated image data processing, usually by running images through different types of computer algorithms, is an overarching goal to reduce the number of images that researchers must manually review. However, differences in camera trap setups and species characteristics can make automatic processing a challenge. Here, we evaluated the detection accuracy and efficiency of a time‐lapse triggered camera trapping technique combined with a pixel change detection algorithm as part of a monitoring program for a translocated population of the rare and federally threatened Louisiana Pinesnake (Pituophis ruthveni). We paired 5 cameras with automated pit tag readers to collect observations of P. ruthveni. We evaluated an image dataset of 1,500,187 images, collected over 7 months, both manually (i.e., researchers looking at each individual picture to determine snake presence) and automatically using a change detection algorithm. There were 18 P. ruthveni observations recorded by the tag readers, 7 of which occurred while a paired camera was not operational. Ten of the tag reader P. ruthveni observations were captured by the paired camera trap, with an additional P. ruthveni observation from a paired camera trap not recorded by the tag reader. There were 132 snake observations of 13 additional species and 18 observations of unknown snakes from the camera traps. The algorithm reduced the number of images reviewers evaluated by an average of 78.5% per camera (range = 37.3–98.7%) but had a 54.5% success rate at detecting observations of P. ruthveni (47.1% for individual images), and a slightly lower 48.9% success rate detecting other large snakes. Large snakes were 4 times more likely to be flagged by the algorithm than small snakes. Our time‐lapse triggered camera trapping technique performed well with respect to P. ruthveni detection accuracy, compared to the tag readers. However, further research is needed to improve quality assurances of camera trap image filtering and object recognition algorithms across different sites or environments.

Impervious surface expansion in the China-Laos border area under the context of geo-economic cooperation: a case study of the Mohan-Boten Economic Cooperation Zone

ABSTRACT Geo-economic cooperation among neighboring countries has promoted landscape changes across their borders, including impervious surface expansion and/or agriculture-forest conversion, which are often presented but usually ignored. Short-revisit multispectral imagery lays a foundation for delineating these gradual or seasonal landscape changes in border regions. China-Laos Mohan-Boten Economic Cooperation Zone (ECZ), which has been launched in March 2016, has recently witnessed a rapid expansion of impervious surfaces. In this study, Sentinel-2 A/B 10 m images over the Mohan-Boten ECZ from 2017 to 2021, and the Continuous Change Detection and Classification algorithm, which combines spatial filtering and temporal consistency check, have been utilized to identify, correct and reconstruct monthly time-series of impervious surface area (ISA). In addition, the patterns and characteristics of impervious surface expansion under the context of China-Laos geo-economic cooperation have been analyzed and the response of impervious surface expansion to geo-economic cooperation has been evaluated. The results show that Mohan-Boten ECZ and China-Laos Railway are two geo-economic contributors to the expansion of impervious surface in both sides of the border. ISA in the ECZ nearly doubled in only five years. ISA in the Boten side expanded much faster than that of the Mohan side in the first half of the study period and exceeded its counterpart for the first time in February 2019. Moreover, the scale and rates of expansion in the dry seasons were much larger in both sides. Three expansion patterns were recognized in the ECZ. This study highlights the potentials of time-series stack of land cover products in quantifying the impacts of geo-economic cooperation on land use change in border areas and in clarifying how border land use change responds to geo-economic cooperation and/or competition.

A Parametric МА-Algorithm for Gaussian Time Series Change Point Detection from Mathematical Expectation

The problem of online detection of a Gaussian time series change connected with a spontaneous abrupt change in its mathematical expectation is solved. For solving the problem, it is proposed to use a parametric controlling algorithm based on the Moving Average method or the MA- algorithm. It is noted that, despite the fact that this algorithm has been known almost since the first works in the field of statistical control, its properties, capabilities, and efficiency in comparison with other change detection algorithms are still (for a number of reasons) have been studied to a very poor extent. The purpose of this work is to thoroughly study the MA-algorithm’s characteristics with the aim to synthesize an optimal time series change detection procedure. The study was carried out using a simulation modeling method. The article presents the structure and element-wise description of the simulation experiment program, which fully replicates the MA-algorithm operation in the online mode. By using the developed program, it has been found that the conventional way of setting the control procedure deciding threshold h, the value of which should provide the preset value of the average time between false alarms, when an alarm signal about the appearance of a change is generated, although in reality the object remains in the "normal" state, is invalid. For a fixed series of the above-mentioned quantities, the correct values of the threshold h are obtained depending on the width of the moving averaging window N of the controlling MA-algorithm. Similarly, the values of the average delay time of producing an alarm signal when a change with the given fixed level  appears. Based on the data obtained, dependences of the control procedure efficiency indicator E on N for a set of different values of average time between false alarms and  have been found. It has been shown that for each such set there is a value of N at which the control procedure is most effective. This optimal procedure is compared with a similar procedure of the cumulative sum algorithm (a CUSUM algorithm), and it has been shown from the comparison results that the MA-algorithm is in general only slightly inferior to the CUSUM algorithm in efficiency, and in some cases even outperforms it.

Adaptive DRL-Based Task Scheduling for Energy-Efficient Cloud Computing

Intelligent task scheduling solutions are highly demanded in the operation of complex cloud data centers so that resources can be utilized in an energy-efficient way while still ensuring various requirements of users. However, the energy problem of task scheduling in cloud environment becomes more challenging with the ever-increasing number of users as well as the constant and unpredictable change of workloads. In this research, we propose an Adaptive Deep Reinforcement Learning-based (ADRL) task scheduling framework for energy-efficient cloud computing. We first present a Change Detection algorithm to detect whether the workload has changed greatly. On this basis, we built an Automatic Generation network to adjust the discount factor of Deep Reinforcement Learning (DRL) dynamically according to the changing workload, which enables faster and more accurate learning. We finally introduce the adaptive DRL to learn the optimal policy of dispatching arriving user requests with the reward aiming to minimize task response time and maximize resource utilization. Simulated experiments have confirmed that the proposed scheduling scheme performs well on accelerating learning convergence and promoting allocation accuracy, thus it is very effective in reducing the average response time of tasks and increasing the CPU utilization rate of resources, which eventually makes the cloud system more energy efficient.

Comparative Analyses of Unsupervised PCA K-Means Change Detection Algorithm from the Viewpoint of Follow-Up Plan

In this study, principal component analysis and k-means clustering (PCAKM) methods for synthetic aperture radar (SAR) data are analyzed to reduce the sensitivity caused by changes in the parameters and input images of the algorithm, increase the accuracy, and make an improvement in the computation time, which are advantageous for scoring in the follow-up plan. Although there are many supervised methods described in the literature, unsupervised methods may be more appropriate in terms of computing time, data scarcity, and explainability in order to supply a trustworthy system. We consider the PCAKM algorithm, which is used as a benchmark method in many studies when making comparisons. Error metrics, computing times, and utility functions are calculated for 22 modified PCAKM regarding difference images and filtering methods. Various images with different characteristics affect the results of the configurations. However, it is evident that the PCAKM becomes less sensitive and more accurate for both the overall results and image results. Scoring by utilizing these results and other map information is a gap and innovation. Obtaining a change map in a fast, explainable, more robust and less sensitive way is one of the aims of our studies on scoring points in the follow-up plan.

Real-time coal classification in thermal power plants

Continuous variations in coal quality have a strong influence on operation of thermal power plants. However, in absence of real-time measurements of quality, the operators remain partially blind to the actual coal being consumed, leading to sub-optimal operation. A two-stage solution is proposed for real-time soft sensing of coal type. First, a novel semi-supervised cascaded clustering algorithm (SSCC) is utilized to extract coal classes from the historical sensor data of the plant and create a coal class library. The online stage includes a coal change detection algorithm to detect transition of coal and a SSCC-based coal classifier that enables real-time classification of coal using only the live sensor data. The algorithms are tested and verified with a set of synthetically generated industrial scale coal mill operation data. The real-time classification of coal can facilitate continuous optimum operation of plant vis-à-vis the emissions, efficiency and maintenance.

Detecting Irrigation Events over Semi-Arid and Temperate Climatic Areas Using Sentinel-1 Data: Case of Several Summer Crops

Irrigation monitoring is of great importance in agricultural water management to guarantee better water use efficiency, especially under changing climatic conditions and water scarcity. This study presents a detailed assessment of the potential of the Sentinel-1 (S1) Synthetic Aperture Radar (SAR) data to detect irrigation events at the plot scale. The potential of the S1 data to detect the irrigation events was carried out using the Irrigation Event Detection Model (IEDM) over semi-arid and temperate oceanic climates in five study sites in south Europe and the Middle East. The IEDM is a decision tree model initially developed to detect irrigation events using the change detection algorithm applied to the S1 time series data. For each study site and at each agricultural plot, all available S1 images during the period of irrigation were used to construct an S1 time series and apply the IEDM. Different types of major summer irrigated crops were analyzed in this study, including Maize, Soybean, Sorghum and Potato, mainly with the sprinkler irrigation technique. The irrigation detection accuracy was evaluated using S1 images and the IEDM against the climatic condition of the studied area, the vegetation development (by means of the normalized difference vegetation index, NDVI) and the revisit time of the S1 sensor. The main results showed generally good overall accuracy for irrigation detection using the S1 data, reaching 67% for all studied sites together. This accuracy varied according to the climatic conditions of the studied area, with the highest accuracy for semi-arid areas and lowest for temperate areas. The analysis of the irrigation detection as a function of the crop type showed that the accuracy of irrigation detection decreases as the vegetation becomes well developed. The main findings demonstrated that the density of the available S1 images in the S1 time series over a given area affects the irrigation detection accuracy, especially for temperate areas. In temperate areas the irrigation detection accuracy decreased from 70% when 15 to 20 S1 images were available per month to reach less than 56% when less than 10 S1 images per month were available over the study sites.

AGVS: A New Change Detection Dataset for Airport Ground Video Surveillance

Change detection is the foundation of intelligent video surveillance of the airport ground. However, experiments have shown that change detection algorithms with good performance on traditional datasets (e.g., CDnet2014) perform poorly in airport ground surveillance. The reason is that traditional datasets focus on the diversity of scenarios, while the practical application requires robustness against various changes in a single scene. We posit that the solution to this problem is to establish a unique dataset for airport ground surveillance and develop specific algorithms for this scenario. In this paper, we present an Airport Ground Video Surveillance benchmark (AGVS) for change detection of the airport ground. AGVS includes 25 long videos, amounting to about 100000 frames and accurate ground truth for all frames. Each video contains multiple challenges specific to the airport ground (e.g., haze, camouflage, strip shape, shadow and illumination change, simultaneous multi-scale objects) and various appearance changes of the aircraft). Change detection ground truth is generated by manual annotation. The AGVS benchmark can be downloaded from <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://www.agvs-caac.com</uri> . Furthermore, we conduct a simple review of current change detection algorithms, both unsupervised or supervised, and then 21 state-of-the-art algorithms are tested and analyzed on the AGVS benchmark. Finally, we conclude with algorithm design principles of change detection for airport ground surveillance.

MAPPING FOREST DISTURBANCE USING PURE FOREST INDEX TIME SERIES AND CCDC ALGORITHM

Abstract. Forest dynamics are closely related to climate change, natural disasters, and ecological diversity. The accumulated Landsat archive provides an unprecedented opportunity for long-term forest dynamics monitoring globally. However, using Landsat time series to detect small-scale and low-intensity disturbance events is still challenging since the moderate spatial resolution of Landsat images and the mixed pixel problem. Towards improving the ability of vegetation index (VI) in characterizing sub-pixel forest dynamics, this paper introduced the spectral mixture analysis (SMA) to develop a novel Pure Forest Index (PFI). The Continuous Change Detection and Classification (CCDC) algorithm was used to detect forest disturbance based on the PFI time series. Cross-comparison shows that PFI is far superior to other conventional VI in indicating forest conditions since it can enhance the spectral signal of the forest and suppress noises from the background. Time series analysis further demonstrates the superiority of PFI in accurately characterizing forest dynamics. The high overall accuracy of 0.96 for the forest disturbance map generated by the proposed approach was achieved. This study highlights a novel VI for accurately tracking subtle forest changes in a heterogeneous landscape.

Ecological quality assessment and monitoring using a time-series remote sensing-based ecological index (ts-RSEI)

ABSTRACT Ecological quality assessment is fundamental to revealing changes in ecological environments before the development of effective ecological conservation policies. The complex ecological environment can be assessed more reliably when multiple remote sensing indices are integrated, such as in the use of the prevalent Remote Sensing-based Ecological Index (RSEI). However, for effective ecological quality assessment, the requirement of acquisition time consistency for images has become an outstanding issue for broadening the application of the RSEI. In this study, we adjusted the RSEI to the Continuous Change Detection and Classification (CCDC) algorithm that predicts synthetic images instead of real images. Based on this algorithm, we mapped time-series RSEI (ts-RSEI), which provide comparable results for tracing the dynamics of ecological quality at any time. Our major findings are as follows: (1) The RSEI is very sensitive to the timespan of the image acquisition dates, with the Mean Absolute Difference (MAD) of 0.111 (19.2%) when the interval between dates exceeds one month. (2) The ts-RSEI from synthetic images is comparable to the RSEI from real images, with the MAD of 0.075 (10.5%), which is superior to that of two real images with the timespan of half-a-month. (3) For Hangzhou, the ecological quality was maintained for almost the past 35 years (the ts-RSEI changed from 0.679 to 0.705). However, special attention should be paid to the spatial polarization between natural (“better”) and human-dominated (“worse”) environments. The high temporal consistency and the capability of any- time mapping of the ts-RSEI are expected to be of value to policy makers and authorities in implementing effective ecological conservation measures.

A Multi-Dimensional Deep Siamese Network for Land Cover Change Detection in Bi-Temporal Hyperspectral Imagery

In this study, an automatic Change Detection (CD) framework based on a multi-dimensional deep Siamese network was proposed for CD in bi-temporal hyperspectral imagery. The proposed method has two main steps: (1) automatic generation of training samples using the Otsu algorithm and the Dynamic Time Wrapping (DTW) predictor, and (2) binary CD using a multidimensional multi-dimensional Convolution Neural Network (CNN). Two bi-temporal hyperspectral datasets of the Hyperion sensor with a variety of land cover classes were used to evaluate the performance of the proposed method. The results were also compared to reference data and two state-of-the-art hyperspectral change detection (HCD) algorithms. It was observed that the proposed method relatively had higher accuracy and lower False Alarm (FA) rate, where the average Overall Accuracy (OA) and Kappa Coefficient (KC) were more than 96% and 0.90, respectively, and the average FA rate was lower than 5%.

Evolution of basic human values orientations: An application of monitoring changes in cluster solutions.

This study enumerates the evolution of basic human values orientations and the dynamic relationship between them, computed from Schwartz's value survey conducted in European nations. For this purpose, eight datasets related to the human value scale were extracted from the European Social Survey; each corresponds to a single round conducted cross-sectionally every two years since 2001. Change detection algorithm was implemented to the cluster solutions of temporal datasets, and the evolution of important clusters was traced. Finding of the study reveals that Universalism and Benevolence values are on the rise in European societies in the last couple of decades. Most of the European inhabitants believe in the smooth group functioning and form the organismic needs of cooperation. The people prefer anxiety-free life, and love for nature, environment, humanity, and kindness to other beings in society are essential constructs for them. They avoid self-centred behaviour and prefer social physiognomies.

Building Change Detection Based on an Edge-Guided Convolutional Neural Network Combined with a Transformer

Change detection extracts change areas in bitemporal remote sensing images, and plays an important role in urban construction and coordination. However, due to image offsets and brightness differences in bitemporal remote sensing images, traditional change detection algorithms often have reduced applicability and accuracy. The development of deep learning-based algorithms has improved their applicability and accuracy; however, existing models use either convolutions or transformers in the feature encoding stage. During feature extraction, local fine features and global features in images cannot always be obtained simultaneously. To address these issues, we propose a novel end-to-end change detection network (EGCTNet) with a fusion encoder (FE) that combines convolutional neural network (CNN) and transformer features. An intermediate decoder (IMD) eliminates global noise introduced during the encoding stage. We noted that ground objects have clearer semantic information and improved edge features. Therefore, we propose an edge detection branch (EDB) that uses object edges to guide mask features. We conducted extensive experiments on the LEVIR-CD and WHU-CD datasets, and EGCTNet exhibits good performance in detecting small and large building objects. On the LEVIR-CD dataset, we obtain F1 and IoU scores of 0.9008 and 0.8295. On the WHU-CD dataset, we obtain F1 and IoU scores of 0.9070 and 0.8298. Experimental results show that our model outperforms several previous change detection methods.

Development of the LCMAP annual land cover product across Hawaiʻi

• Generated several training datasets by integrating existing datasets. • Produced annual land cover maps for the Hawaiian Islands at 30-m resolution over 20 years with different training datasets and Change detection and classification (CCDC) algorithm. • Summarized the effect of training datasets (e.g., spatial coverage, quality) on accuracies for different land cover types. • Revealed annual land cover changes in Hawaiʻi. Following the completion of land cover and change (LCC) products for the conterminous United States (CONUS), the U.S. Geological Survey's (USGS’s) Land Change Monitoring, Assessment, and Projection initiative has broadened the capability of characterizing continuous historical land change across the full Landsat records for Hawaiʻi at 30-meter resolution. One of the challenges of implementing the LCMAP framework to process annual land cover maps in Hawaiʻi is to collect sufficient high-quality training data. Although multiple datasets depicting land cover information are available in Hawaiʻi, they covered limited time frames and were produced from various remote sensing sources with different, classification categories, spatial resolution, and mapping accuracies. No solo product is suitable to provide LCMAP training data labels on its own. In this paper, we focused on enhancing the LCMAP training datasets to generate land cover products from 2000 to 2019 in Hawaiʻi. A total of 200 independent reference data plots were generated and manually interpreted for validating the mapping results produced by the training datasets. The results revealed that using the appropriate filter of multiple products as training data pools improved the classification model performance. The effect of training datasets (e.g., spatial coverage, quality) on accuracies for different land cover types were summarized. The LCMAP land surface change products for Hawaiʻi are available at https://doi.org/10.5066/P91E8M23 .

Object-Based Change Detection Algorithm with a Spatial AI Stereo Camera.

This paper presents a real-time object-based 3D change detection method that is built around the concept of semantic object maps. The algorithm is able to maintain an object-oriented metric-semantic map of the environment and can detect object-level changes between consecutive patrol routes. The proposed 3D change detection method exploits the capabilities of the novel ZED 2 stereo camera, which integrates stereo vision and artificial intelligence (AI) to enable the development of spatial AI applications. To design the change detection algorithm and set its parameters, an extensive evaluation of the ZED 2 camera was carried out with respect to depth accuracy and consistency, visual tracking and relocalization accuracy and object detection performance. The outcomes of these findings are reported in the paper. Moreover, the utility of the proposed object-based 3D change detection is shown in real-world indoor and outdoor experiments.

Real-Time Software for the Efficient Generation of the Clumping Index and Its Application Based on the Google Earth Engine

Canopy clumping index (CI) is a key structural parameter related to vegetation phenology and the absorption of radiation, and it is usually retrieved from remote sensing data based on an empirical relationship with the Normalized Difference between Hotspot and Darkspot (NDHD) index. A rapid production software was developed to implement the CI algorithm based on the Google Earth Engine (GEE) to update current CI products and promote the application of CI in different fields. Daily, monthly, and yearly global CI products are continuously generated and updated in real-time by the software. Users can directly download the product or work with CI without paying attention to data generation. For the application case study, a change detection algorithm, LandTrendr, was implemented on the GEE to examine the global CI trend from 2000 to 2020. The results indicate that the area of increase trend (28.7%, ΔCI &gt; 0.02) is greater than that of the decrease trend (17.1%, ΔCI &lt; −0.02). Our work contributes toward the retrieval, application, and validation of CI.

A high-resolution feature difference attention network for the application of building change detection

Deep learning based change detection has brought a significant improvement in the accuracy and efficiency when compared with conventional machine learning methods. However, the issues of the lack of differential information and the diversity of the scale features of artificial objects are crucial barriers to the application of building change detection algorithms. A novel deep learning based approach named the high-resolution feature difference attention network (HDANet) is proposed in this work to solve these issues. HDANet can handle the change characteristics well, due to the Siamese network structure. To tackle the loss of the spatial features of buildings caused by the multiple successive down-sampling operations in the current change detection algorithms using fully convolutional networks (FCNs), a multi-resolution parallel structure is introduced in HDANet, and the image information with different resolutions is comprehensively employed, without any spatial information loss. Moreover, an innovative difference attention module is elaborated for the enhancement of the sensitivity to difference information, to keep the building change information. The experimental results obtained on building change detection datasets confirm that HDANet can improve the differential feature representation for change detection, and the performance of the building change detection is also superior to that of the other advanced change detection methods.

Contrasting Forest Loss and Gain Patterns in Subtropical China Detected Using an Integrated LandTrendr and Machine-Learning Method

China has implemented a series of forestry law, policies, regulations, and afforestation projects since the 1970s. However, their impacts on the spatial and temporal patterns of forests have not been fully assessed yet. The lack of an accurate, high-resolution, and long-term forest disturbance and recovery dataset has impeded this assessment. Here we improved the forest loss and gain detections by integrating the LandTrendr change detection algorithm with the Random Forest (RF) machine-learning method and applied it to assess forest loss and gain patterns in the Zhejiang, Jiangxi, and Guangxi Provinces of the subtropical vegetation in China. The accuracy evaluation indicated that our approach can adequately detect the spatial and temporal distribution patterns in forest gain and loss, with an overall accuracy of 93% and the Kappa coefficient of 0.89. The forest loss area was 8.30 × 104 km2 in the Zhejiang, Jiangxi, and Guangxi Provinces during 1986–2019, accounting for 43.52% of total forest area in 1986, while the forest gain area was 20.25 × 104 km2, accounting for 106.19% of total forest area in 1986. Although the interannual variation patterns were similar among three provinces, the forest loss and gain area and the magnitude of change trends were significantly different. Guangxi has the largest forest loss and gain area and increasing trends, followed by Jiangxi, and the least in Zhejiang. The variations in annual forest loss and gain area can be mostly explained by the timelines of major forestry policies and regulations. Our study would provide an applicable method and data for assessing the impacts of forest disturbance events and forestry policies and regulations on the spatial and temporal patterns of forest loss and gain in China, and further contributing to regional and national forest carbon and greenhouse gases budget estimations.

A Temporal-Reliable Method for Change Detection in High-Resolution Bi-Temporal Remote Sensing Images

Very-high-resolution (VHR) bi-temporal images change detection (CD) is a basic remote sensing images (RSIs) processing task. Recently, deep convolutional neural networks (DCNNs) have shown great feature representation abilities in computer vision tasks and have achieved remarkable breakthroughs in automatic CD. However, a great majority of the existing fusion-based CD methods pay no attention to the definition of CD, so they can only detect one-way changes. Therefore, we propose a new temporal reliable change detection (TRCD) algorithm to solve this drawback of fusion-based methods. Specifically, a potential and effective algorithm is proposed for learning temporal-reliable features for CD, which is achieved by designing a novel objective function. Unlike the traditional CD objective function, we impose a regular term in the objective function, which aims to enforce the extracted features before and after exchanging sequences of bi-temporal images that are similar to each other. In addition, our backbone architecture is designed based on a high-resolution network. The captured features are semantically richer and more spatially precise, which can improve the performance for small region changes. Comprehensive experimental results on two public datasets demonstrate that the proposed method is more advanced than other state-of-the-art (SOTA) methods, and our proposed objective function shows great potential.