High IoU Research Articles

TSRDet: A Table Structure Recognition Method Based on Row-Column Detection

As one of the most commonly used and important data carriers, tables have the advantages of high structuring, strong readability and strong flexibility. However, in reality, tables usually present various forms, such as Excel, images, etc. Among them, the information in the table image cannot be read directly, let alone further applied. Therefore, the research related to image-based table recognition is crucial. It contains the table structure recognition and the table content recognition. Among them, table structure recognition is the most important and difficult task because the table structure is abstract and changeable. In order to address this problem, we propose an innovative table structure recognition method, named TSRDet (Table Structure Recognition based on object Detection). It includes a row-column detection method, named SACNet (StripAttention-CenterNet) and the corresponding post-processing. SACNet is an improved version of the original CenterNet. The specific improvements include the following: firstly, we introduce the Swin Transformer as the encoder to obtain the global feature map of the image. Then, we propose a plug-and-play row-column attention module, including a channel attention module and a row-column spatial attention module. It improves the detection accuracy of rows and columns by capturing long-range row-column feature maps in the image. After completing the row-column detection, this paper also designs a simple and fast post-processing to generate the table structure based on the row-column detection results. Experimental results show that for row-column detection, SACNet has high detection accuracy, even at a high IoU threshold. Specifically, when the threshold is 0.75, its mAP of row detection and column detection still exceeds 90%, which is 91.40% and 92.73% respectively. In addition, in the comparative experiment with the existing object detection methods, SACNet’s performance was significantly better than that of all others. For table structure recognition, the TEDS-Struct score of TSRDet is 95.7%, which shows competitive performance in table structure recognition, and verifies the rationality and superiority of the proposed method.

Wildlife target detection based on improved YOLOX-s network

To addresse the problem of poor detection accuracy or even false detection of wildlife caused by rainy environment at night. In this paper, a wildlife target detection algorithm based on improved YOLOX-s network is proposed. Our algorithm comprises the MobileViT-Pooling module, the Dynamic Head module, and the Focal-IoU module.First, the MobileViT-Pooling module is introduced.It is based on the MobileViT attention mechanism, which uses a spatial pooling operator with no parameters as a token mixer module to reduce the number of network parameters. This module performs feature extraction on three feature layers of the backbone network output respectively, senses the global information and strengthens the weight of the effective information. Second, the Dynamic Head module is used on the downstream task of network detection, which fuses the information of scale sensing, spatial sensing, and task sensing and improves the representation ability of the target detection head. Lastly, the Focal idea is utilized to improve the IoU loss function, which balances the learning of high and low quality IoU for the network. Experimental results reveal that our algorithm achieves a notable performance boost with mAP@0.5 reaching 87.8% (an improvement of 7.9%) and mAP@0.5:0.95 reaching 62.0% (an improvement of 5.3%). This advancement significantly augments the night-time wildlife detection accuracy under rainy conditions, concurrently diminishing false detections in such challenging environments.

Multiresolution cascaded attention U-Net for localization and segmentation of optic disc and fovea in fundus images

Identification of retinal diseases in automated screening methods, such as those used in clinical settings or computer-aided diagnosis, usually depends on the localization and segmentation of the Optic Disc (OD) and fovea. However, this task is difficult since these anatomical features have irregular spatial, texture, and shape characteristics, limited sample sizes, and domain shifts due to different data distributions across datasets. This study proposes a novel Multiresolution Cascaded Attention U-Net (MCAU-Net) model that addresses these problems by optimally balancing receptive field size and computational efficiency. The MCAU-Net utilizes two skip connections to accurately localize and segment the OD and fovea in fundus images. We incorporated a Multiresolution Wavelet Pooling Module (MWPM) into the CNN at each stage of U-Net input to compensate for spatial information loss. Additionally, we integrated a cascaded connection of the spatial and channel attentions as a skip connection in MCAU-Net to concentrate precisely on the target object and improve model convergence for segmenting and localizing OD and fovea centers. The proposed model has a low parameter count of 0.8 million, improving computational efficiency and reducing the risk of overfitting. For OD segmentation, the MCAU-Net achieves high IoU values of 0.9771, 0.945, and 0.946 for the DRISHTI-GS, DRIONS-DB, and IDRiD datasets, respectively, outperforming previous results for all three datasets. For the IDRiD dataset, the MCAU-Net locates the OD center with an Euclidean Distance (ED) of 16.90 pixels and the fovea center with an ED of 33.45 pixels, demonstrating its effectiveness in overcoming the common limitations of state-of-the-art methods.

UAV detection technology research based on deep learning

Abstract In the process of UAV target recognition, the target is small, the feature is not obvious, and the recognition accuracy is low, a UAV target detection method called YOLOv8-UD is proposed based on YOLOv8. This method replaces downsampling layers with SPD convolutions to mitigate fine-grained feature loss issues. It employs a BiFPN structure to get information which is multi-scale, enhancing representation of global semantic message and Improve awareness of small goals. As a result, to target detection, it significantly enhances the correctness and effectiveness. Additionally, to address difficulty of detecting low-resolution small targets, the network structure was modified by adding a detection head, for small target, thereby enhancing capability to detect it. Combining an improved loss function based on Inner-IoU accelerated model convergence and improved detection accuracy of samples at high IoU thresholds. We can tell by experiments that The algorithm proposed in this paper have 0.805 and 0.334 in mAP(0.5) and mAP(0.5:0.95) metrics, respectively. Compared to YOLOv8, improvements of 4.6% and 7% are achieved in mAP(0.5) and mAP (0.5:0.95) metrics. The experiments validate that the proposed way significantly enhances accuracy to UAV small target detection.

Predicting nitrous oxide contaminants in Cauvery basin using region-based convolutional neural network

Nitrous oxide (N2O) in riverbeds affects hydrological processes by contributing to the greenhouse effect, indicating poor water quality, disrupting biogeochemical cycling, and linking to eutrophication. Elevated N2O levels signal environmental issues, impacting aquatic life and necessitating precise forecasting for effective environmental management and reduced greenhouse gas emissions. Precisely forecasting nitrous oxide (N2O) emissions from riverbeds is paramount for effective environmental management, given its significant potency as a greenhouse gas. This study focuses on the difficulties related to spatial feature extraction and modeling accuracy in predicting N2O in riverbeds in Tamil Nadu. To address the obstacles, the research suggests utilizing the Deep Learning Based Prediction of Nitrous Oxide Contaminants (DL-PNOC), which studies the N2O contaminants in water using Region-based Convolutional Neural Network (RCNN) for spatial feature extraction, to predict nitrous oxide contaminants. The study is centered on the Cauvery River Basin located in Tamil Nadu, where the emission of N2O is a matter of environment. The outcomes encompass the specialized N2O contaminant model for riverbeds and the implementation of RCNN achieves precise N2O forecasting. The DL-PNOC approach combines a contaminant model with RCNN deep learning techniques to capture spatial characteristics and predict N2O pollutants accurately. Furthermore, using the River Bed Dynamics Simulator reinforces the dependability of the findings. The DL-PNOC approach has exhibited encouraging results, as evidenced by the following metrics: a high IoU of 88.66%, precision of 88.96%, recall of 90.03%, F1 score of 89.22%, and low RMSE and MAE values of 9.14% and 7.59%, respectively. The findings highlight the efficacy of the DL-PNOC approach in precisely forecasting N2O pollutants in river sediments.

Segmentation of liver CT images based on weighted medical transformer model.

Deep convolutional neural networks have made significant strides in the field of medical image segmentation. Although existing convolutional structures enhance performance by leveraging local image information, they often lose the interdependence information between contexts. Therefore, the article utilizes the multi-attention mechanism of the Transformer structure to more comprehensively express relationships between contexts and introduced the Transformer network architecture into the field of medical image segmentation. Most models based on this Transformer structure typically require large datasets for training. However, in the medical field, the limited size of datasets makes training models with the Transformer structure challenging. To address this, the article propose a Weighted Medical Transformer (WMT) model that imposes low requirements on dataset quantity. The weighting mechanism in the WMT model aims to improve the issue of inaccurate relative positional coding when dealing with small medical datasets. Additionally, a coarse-grained and fine-grained segmentation mechanism is introduced, focusing on both the detailed aspects within image blocks and the boundary information connecting blocks. Experimental results on a liver dataset demonstrate that the model achieves F1 and IoU scores of 88.48% and 79.41%, respectively. Results on the MoNuSeg dataset show comparable high F1 and IoU scores of 79.58% and 66.19%, respectively. The model's accuracy surpasses that of U-Net++ and U-Net models. Compared to other models, this approach is applicable to scenarios with limited datasets, exhibiting high execution efficiency and accuracy.

Learning Reliable Dense Pseudo-Labels for Point-Level Weakly-Supervised Action Localization

Point-level weakly-supervised temporal action localization aims to accurately recognize and localize action segments in untrimmed videos, using only point-level annotations during training. Current methods primarily focus on mining sparse pseudo-labels and generating dense pseudo-labels. However, due to the sparsity of point-level labels and the impact of scene information on action representations, the reliability of dense pseudo-label methods still remains an issue. In this paper, we propose a point-level weakly-supervised temporal action localization method based on local representation enhancement and global temporal optimization. This method comprises two modules that enhance the representation capacity of action features and improve the reliability of class activation sequence classification, thereby enhancing the reliability of dense pseudo-labels and strengthening the model’s capability for completeness learning. Specifically, we first generate representative features of actions using pseudo-label feature and calculate weights based on the feature similarity between representative features of actions and segments features to adjust class activation sequence. Additionally, we maintain the fixed-length queues for annotated segments and design a action contrastive learning framework between videos. The experimental results demonstrate that our modules indeed enhance the model’s capability for comprehensive learning, particularly achieving state-of-the-art results at high IoU thresholds.

Semantic segmentation of urban environments: Leveraging U-Net deep learning model for cityscape image analysis.

Semantic segmentation of cityscapes via deep learning is an essential and game-changing research topic that offers a more nuanced comprehension of urban landscapes. Deep learning techniques tackle urban complexity and diversity, which unlocks a broad range of applications. These include urban planning, transportation management, autonomous driving, and smart city efforts. Through rich context and insights, semantic segmentation helps decision-makers and stakeholders make educated decisions for sustainable and effective urban development. This study investigates an in-depth exploration of cityscape image segmentation using the U-Net deep learning model. The proposed U-Net architecture comprises an encoder and decoder structure. The encoder uses convolutional layers and down sampling to extract hierarchical information from input images. Each down sample step reduces spatial dimensions, and increases feature depth, aiding context acquisition. Batch normalization and dropout layers stabilize models and prevent overfitting during encoding. The decoder reconstructs higher-resolution feature maps using "UpSampling2D" layers. Through extensive experimentation and evaluation of the Cityscapes dataset, this study demonstrates the effectiveness of the U-Net model in achieving state-of-the-art results in image segmentation. The results clearly shown that, the proposed model has high accuracy, mean IOU and mean DICE compared to existing models.

An oil and gas pipeline inspection UAV based on improved YOLOv7

This study proposes a method of autonomous navigation UAV for oil and gas pipeline (OGP) dial detection based on the improved YOLOv7 model. The canny edge detection algorithm is applied in identifying the edges of the pipeline, and the Hough transform algorithm is used to detect the pipeline in a straight line. The intelligent UAV P600 is guided to patrol the oil and gas dials (OGD) along the pipeline, and the trained improved YOLOv7-based model is adopted to identify the OGD data. Dial recognition is divided into two stages, that is, dial contour detection and dial reading recognition. For the dial recognition rate (RR), the Levenstein distance, a commonly used method, is introduced, thereby calculating the distance between two character sequences. Meanwhile, an integrated global attention mechanism (GAM) is proposed based on the YOLOv7 model, aiming at extracting more informative features. With this mechanism, the channel and spatial aspects of the features are effectively captured, and the importance of cross-dimensional interactions is increased. By introducing GAM attention mechanism in the backbone and head of YOLOv7, the network’s ability in efficiently extracting depth and primary features is enhanced. ACmix (a hybrid model combining the advantages of self-attentiveness and convolution) is also included, with ACmix module improved. The improved ACmix module has the objectives of enhancing feature extraction capability of backbone network and accelerating network convergence. By substituting 3 × 3 convolutional block with 3 × 3 ACmixBlock and adding a jump connection and a 1 × 1 convolutional structure between the ACmixBlock modules, E-ELAN module in YOLOv7 network is also improved, thus optimizing E-ELAN network, enriching features extracted by E-ELAN network, and reducing inference time of YOLOv7 model. As indicated by comparing the experimental results of the six model algorithms (improved YOLOv7, YOLOv7, YOLOX, YOLOv5, YOLOv6 and Faster R-CNN), the improved YOLOv7 model has higher mAP, faster RR, faster network convergence, and higher IOU. In addition, a generic real dataset, called custom dial reading dataset, is presented. With well-defined evaluation protocol, this dataset allows for a fair comparison of various methods in future work.

Combination of Image Enhancement and U-Net Architecture for Cervical Cell Semantic Segmentation

Cervical cancer is the second leading cause of death in women and ranks fourth as a disease that occurs in women worldwide. Cervical cancer is a disease that is difficult to detect and can be detected when it is in an advanced stage. This requires early prevention by carrying out a pap-smear examination. Pap-smear examination manually requires a relatively long time, so a tool is needed by segmentation. Segmentation is image processing by performing perfection between the intended object and the background. One of the CNN methods commonly used in medical image segmentation is the U-Net architecture. Segmentation in this study was carried out on the nucleus and cytoplasm of the Herlev dataset using the U-Net architecture combined with data augmentation and image enhancement. In the learning process, this research resulted in a fairly high IoU value of 78% and an RMSE close to 20%. The results of this study also yielded an accuracy value of 89%, with an average precision, recall and F1 score of 89%, 89% and 88.67%, respectively. This shows that the combination of the CNN U-Net architecture with image quality improvement and data augmentation is quite good at segmenting cervical cells for the nucleus and cytoplasm

Automatic Labeling of Natural Landmarks for Wheelchair Motion Planning

Labeling landmarks for the mobile plan of the automatic electric wheelchair is essential, because it can assist disabled people. In particular, labeled landmark images will help the wheelchairs to locate landmarks and move more accurately and safely. Here, we propose an automatic detection of natural landmarks in RGBD images for navigation of mobile platforms in an indoor environment. This method can reduce the time for manually collecting and creating a dataset of landmarks. The wheelchair, equipped with a camera system, is allowed to move along corridors to detect and label natural landmarks automatically. These landmarks contain the camera and wheelchair positions with the 3D coordinates when storing the labeled landmark. The feature density method is comprised of Oriented FAST and Rotated BRIEF (ORB) feature extractors. Moreover, the central coordinates of the marked points in the obtained RGB images will be mapped to the images with the depth axis for determining the position of the RGB-D camera system in the spatial domain. An encoder and kinematics equations are applied to determine the position during movement. As expected, the system shows good results, such as a high IoU value of over 0.8 at a distance of less than 2 m and a fast time of 41.66 ms for object detection. This means that our technique is very effective for the automatic movement of the wheelchair.

Near Real-Time Flood Mapping with Weakly Supervised Machine Learning

Advances in deep learning and computer vision are making significant contributions to flood mapping, particularly when integrated with remotely sensed data. Although existing supervised methods, especially deep convolutional neural networks, have proved to be effective, they require intensive manual labeling of flooded pixels to train a multi-layer deep neural network that learns abstract semantic features of the input data. This research introduces a novel weakly supervised approach for pixel-wise flood mapping by leveraging multi-temporal remote sensing imagery and image processing techniques (e.g., Normalized Difference Water Index and edge detection) to create weakly labeled data. Using these weakly labeled data, a bi-temporal U-Net model is then proposed and trained for flood detection without the need for time-consuming and labor-intensive human annotations. Using floods from Hurricanes Florence and Harvey as case studies, we evaluated the performance of the proposed bi-temporal U-Net model and baseline models, such as decision tree, random forest, gradient boost, and adaptive boosting classifiers. To assess the effectiveness of our approach, we conducted a comprehensive assessment that (1) covered multiple test sites with varying degrees of urbanization, and (2) utilized both bi-temporal (i.e., pre- and post-flood) and uni-temporal (i.e., only post-flood) input. The experimental results showed that the proposed framework of weakly labeled data generation and the bi-temporal U-Net could produce near real-time urban flood maps with consistently high precision, recall, f1 score, IoU score, and overall accuracy compared with baseline machine learning algorithms.

Apple object detection based on improved YOLOX

To enhance the working efficiency of apple picking systems by improving the accuracy of object detection in natural scenes, an improved apple detection network based on YOLOX-s is proposed. The self-attention residual module is added to the last layer of the improved YOLOX-s, which is used to add global feature information for small apples. An additional object detection head is added in YOLOX-s to strengthen the feature information of dense objects. Extra convolutional branches are added into the spatial pyramid pooling structure to enhance feature fusion and compensate for lost object location information. The loss function of the network is replaced by the α-CIoU loss, which is used to improve the bbox regression accuracy by up-weighting the loss and gradient of the high IoU prediction box. Experiment results show that the mAP50 value and the recall rate of the improved network reached 91.8% and 97%, respectively. Therefore, the improved network is superior to the existing detection networks in detection accuracy. Meanwhile, the detection time increased slightly, but it still meets the requirements of real-time detection of the picking system.

Detection and quantification of cotton trichomes by deep learning algorithm

Cotton is one of the most economically important agricultural crops with the harvest of cotton fibers. Trichomes are initiated from the plant epidermis, which is responsible for plant resistance to many abiotic and biotic stresses. Cotton fibers share similar underlying developmental molecular mechanisms with trichomes on the stem and leaf margins of cotton. Compared with fibers, it is easier to observe the trichome phenotype of cotton stem and leaf margins. However, because of the variability in coloration and the high occlusion of the trichomes, counting the density and observing the distribution of trichomes is still highly laborious. To help resolve this problem, the present research provides a solution based on deep learning in the screening of cotton trichomes, which makes it possible to monitor the phenotype of the cotton efficiently. With a fluorescent microscopy imaging procedure, we clearly observed that there are two types of trichome phenotype inGossypium hirsutum(‘TM-1′), which are singled and clustered trichomes. The clustered trichome originates from two or more adjacent epidermal cells with no branches. The TM1 trichome was used to create a collection of singled and clustered trichome pictures. Following data production, we evaluated the effectiveness of several you only look once (YOLO) model iterations in the context of identifying singled and clustered trichomes, respectively. Results show that the YOLOv5s model outperforms YOLOv3 and YOLOv4 in average accuracy, recall, accuracy, and F1 score. We utilized the clustered trichome data set to assess the performance of YOLOv5s and Mask R-CNN in order to more precisely detect the real trichome distribution with simultaneous detection of single and clustered trichomes. The obtained findings show that both models may be utilized to concurrently identify singled and clustered trichomes since the discrepancy between the detected values of the two models and the real values is quite minimal. Mask R-CNN has a high IOU value, whereas YOLOv5s has a low false-positive detection rate and fast detection speed. Our models and datasets can be used as a starting point to measure the trichome trait in other cotton species and establish a cotton trichome detection and counting system based on deep learning technology.

Lung lobe segmentation in computed tomography images based on multi‐feature fusion and ensemble learning framework

AbstractLung lobe segmentation in computed tomography (CT) images can be regarded as essential supporting information for the diagnosis and treatment of lung diseases, yet it is a challenging uncertainty for the complex segmentation task due to the diverse structures like indistinguishable pulmonary arteries and veins, unpredictable pathological deformation and blurring pulmonary fissures. To circumvent these challenges, we present a productive deep learning network based on multi‐feature fusion and ensemble learning approach to highlight pulmonary fissure representation and suppress other structures for lung lobe segmentation in CT images. Considering that pulmonary fissures are the physical boundaries of lung lobes, a multi‐feature fusion approach is presented to integrate original images, enhanced fissure magnitude images, and enhanced fissure orientation images to highlight shape representation for semantic segmentation. In addition, a deep‐supervised ensemble learning network is employed to combine multiple inducers for lung lobe segmentation improvement. The performance of the proposed deep learning model is validated in experiments with an available dataset, it acquired a high IoU value and a low Hausdorff distance compared with manual references. Experimental results demonstrate that the proposed computational model based on multi‐feature fusion and deep‐supervised ensemble learning framework has an improved predictive performance than many state‐of‐the‐art deep neural networks in lung lobe segmentation.

Architecture of Deep Convolutional Encoder-Decoder Networks for Building Footprint Semantic Segmentation

Building extraction from high-resolution aerial images is critical in geospatial applications such as telecommunications, dynamic urban monitoring, updating geographic databases, urban planning, disaster monitoring, and navigation. Automatic building extraction is a massive task because buildings in various places have varied spectral and geometric qualities. As a result, traditional image processing approaches are insufficient for autonomous building extraction from high-resolution aerial imaging applications. Automatic object extraction from high-resolution images has been achieved using semantic segmentation and deep learning models, which have become increasingly important in recent years. In this study, the U-Net model was used for building extraction, initially designed for biomedical image analysis. The encoder part of the U-Net model has been improved with ResNet50, VGG19, VGG16, DenseNet169, and Xception. However, three other models have been implemented to test the performance of the model studied: PSPNet, FPN, and LinkNet. The performance analysis through the intersection of union method has shown that U-Net with the VGG16 encoder presents the best results compared to the other models with a high IoU score of 83.06%. This research aims to examine the effectiveness of these four approaches for extracting buildings from high-resolution aerial data.

Facial expression recognition method based on PSA-YOLO network.

In order to improve the recognition speed and accuracy of face expression recognition, we propose a face expression recognition method based on PSA-YOLO (Pyramids Squeeze Attention-You Only Look Once). Based on CSPDarknet53, the Focus structure and pyramid compression channel attention mechanism are integrated, and the network depth reduction strategy is adopted to build a PSA-CSPDarknet-1 lightweight backbone network with small parameters and high accuracy, which improves the speed of face expression recognition. Secondly, in the neck of the network, a spatial pyramid convolutional pooling module is built, which enhances the spatial information extraction ability of deep feature maps with a very small computational cost, and uses the α-CIoU loss function as the bounding box loss function to improve the recognition accuracy of the network for targets under high IoU threshold and improve the accuracy of face expression recognition. The proposed method is validated on the JAFFE, CK+, and Cohn-Kanade datasets. The experimental results show that the running time of the proposed method and the comparison method is reduced from 1,800 to 200 ms, and the recognition accuracy is increased by 3.11, 2.58, and 3.91%, respectively, so the method proposed in this paper has good applicability.

Segmentation of Sella Turcica in X-ray Image based on U-Net Architecture

The computer-aided evaluation of medical imaging and orthodontic examinations requires the segmentation of Sella Turcica for feature extraction using cephalometric X-ray images, which is a novel and complex task. However, x-ray image segmentation is more complicated than CT and MR images due to less association of dense thin white tissue of Sella Turcica that is difficult to distinguish from cephalometric radiographs. This paper presents deep learning methods for completely automated Sella Turcica segmentation from cephalometric X-ray images using a fully connected neural network ("U-Net"). The experiments were performed and analyzed using secondary datasets provided by PGIMER, Chandigarh. The different pre-trained U-Net models (VGG16 and 19, ResNet34, 50, and 152) are used as an encoder to improve the prediction accuracy of the Sella structure. The author compared the IoU score of all the considered pre-trained models and found that VGG19 and ResNet34 have high IoU scores compared to other models, but out of all the considered models, ResNet34 gives good prediction results.

Cascaded multi-point regression Network for high-quality generic lesion detection

With such a high incidence of cancer, research on high-quality lesion detection is extremely significant to improve diagnostic efficiency. With the ability to correctly recognize the lesion, it is even more significant to accurately locate the lesion, which can contribute effectively to the subsequent segmentation and radiation of the lesion. In this paper, we propose a cascaded multi-point regression detection network, which can predict offsets of multiple local points of proposals. Different from traditional regression and location based on keypoints, Multi-point regression (MPR) predicts multi offsets for local points that are position sensitive instead of regressing the whole object proposal. In order to reduce the influence of background region on the final box regression, we set the corresponding regression weight for each local point according to different IoU thresholds. At low IoU threshold, the category score is used as the regression weight for each point. At high IoU threshold, a binary weight prediction is proposed to filter out the background area. For improving the transmission and extraction of the underlying geometric location information, we build a bottom-up information propagation path in FPN as an augmented feature pyramid network (AFPN). We use the dataset DeepLesion to evaluate our network, and the experimental results show that our method can perform as expected. The detection network can improve the quality of lesion detection in terms of localization accuracy compared to most mainstream models.

Multibranch Unsupervised Domain Adaptation Network for Cross Multidomain Orchard Area Segmentation

Although unsupervised domain adaptation (UDA) has been extensively studied in remote sensing image segmentation tasks, most UDA models are designed based on single-target domain settings. Large-scale remote sensing images often have multiple target domains in practical applications, and the simple extension of single-target UDA models to multiple target domains is unstable and costly. Multi-target unsupervised domain adaptation (MTUDA) is a more practical scenario that has great potential for solving the problem of crossing multiple domains in remote sensing images. However, existing MTUDA models neglect to learn and control the private features of the target domain, leading to missing information and negative migration. To solve these problems, this paper proposes a multibranch unsupervised domain adaptation network (MBUDA) for orchard area segmentation. The multibranch framework aligns multiple domain features, while preventing private features from interfering with training. We introduce multiple ancillary classifiers to help the model learn more robust latent target domain data representations. Additionally, we propose an adaptation enhanced learning strategy to reduce the distribution gaps further and enhance the adaptation effect. To evaluate the proposed method, this paper utilizes two settings with different numbers of target domains. On average, the proposed method achieves a high IoU gain of 7.47% over the baseline (single-target UDA), reducing costs and ensuring segmentation model performance in multiple target domains.