Single-stage object detector with local binary pattern for remote sensing images

Lightweight object detection and recognition models are extremely crucial for in-orbit applications, which is the most critical factor for whether deep learning-based object detection and recognition algorithms can be applied to remote sensing satellites for real-time or near real-time processing. Global information is extremely important for object detection and recognition of remote sensing images. However, due to the high computational cost, the existing CNN-based lightweight models over-emphasize on the extraction of local information, while ignoring the global information. For this reason, we propose a lightweight object detection and recognition model (Lightweight Global-Local Detection, LGLDet) based on the especially light global modeling structure. In LGLDet, light global-local module (LGLM) is proposed to extract the global and local information. The LGLM consists of Point2Patch Non-Local (P2PNL), local branch and skip connection. Specifically, P2PNL is proposed to reduce the computation of global long-range dependency modeling. In addition, the feature fusion part and detection head are also designed in a lightweight way. In the experiments, the proposed method can achieve optimal performance with fewer parameters and lower computational complexity than existing CNN-based lightweight models and transformer-based lightweight models with similar parameters or computational complexity. The code will be released on the site of https://github.com/dyl96/LGLDet.

In recent years, computer vision technology has developed rapidly, as one of its important research directions, object detection has received widespread attention due to its high accuracy. Meanwhile, object detection has many application fields, such as intelligent transportation, medical and health, security systems, etc. Traditional object detection methods have limitations when applied to complex real-world scenarios. To improve the shortcomings of conventional methods, deep learning based object detection algorithms have significantly improved the efficiency of object detection and become a research hotspot in object detection. This article summarizes the algorithms into two-stage and one-stage object detection algorithms based on the technical processes and structural differences in handling object detection tasks. Firstly, several common two-stage and one-stage object detection algorithms and their applications in real-world scenarios are introduced. Then their data sets and algorithm performance are analyzed and compared. Finally, according to the results of the comparison, the existing problems of the two-stage object detection algorithm and the one-stage object detection method are discussed, and their future development directions are pointed out.

Object detection is increasingly in demand in IoT service applications. Deep learning based object detection algorithms are now in fashion. As the most popular multi-scale object detection network at present, Feature Pyramid Network achieves feature augmentation by fusing features of neighboring layers. It is widely used in the most advanced object detectors to detect objects of different scales. In this paper, we propose a new attention mechanism guided bidirectional feature pyramid architecture named AgBFPN to enhance the transfer of semantic and spatial information between each feature map. We design Channel Attention Guided Fusion(CAGF) Module and Spatial Attention Guided Fusion(SAGF) Module to enhance feature fusion. The CAGF mitigates the loss of information induced by channel reduction and better transfers the semantic information from high-level to low-level features. The SAGF passes the rich spatial information of shallow features into deep features. Our experiments show that AgBFPN achieves higher Average Precision for multi-scale object detection.

The article discusses a texture operator called Local Binary Patterns (LBP) and its applications in image processing and object detection. We provide a description of the algorithm for computing LBP together with a rationale for using LBP as a feature for object detection and image recognition. Based on the algorithm we show that LBP features have a low computational overhead compared to more complicated image features such as the commonly used SIFT or SURF features or neural network based approaches because they exploit the use of extremely fast bitwise and integer operations of the CPU. We demonstrate that LBP is robust to changes in brightness, contrast, image rotation, image scale. We develop two enhancements for LBP that improve its resistance to camera noise and enhance the discriminative power of LBP when it is used as a feature for machine learning algorithms. We present the results on a challenging real-world object detection task. Keywords: computer vision, object detection, local binary patterns.

Monitoring the spatial distribution of tailing ponds is of great significance for environmental governance in mining areas. Remote sensing technology is an effective tool to quickly obtain the location information of tailing ponds in large areas. However, most current studies about tailing pond detection focus on using images from single sensor and are with relatively low detection accuracy in complex backgrounds. In this paper, we present a framework to annotate the location of Tailing Ponds (TPs) from multi-source and high-resolution remote sensing images. In the proposed framework, an improved You Only Look Once (YOLO) deep learning based object detection model is employed, which embeds a high-density feature aggregation module, a low-parameter feature aggregation module and a refined loss function. The introductions of two new modules can make the network better capture global information about the object and accelerate the detection speed, while the introduction of a refined loss function can make the target box regression process more robust and avoiding divergence during training. To train reliable and robust YOLO based TP detection model, sufficient training samples in different backgrounds were collected, and some different sample augmentation methods were also employed. Experimental results validate that, the proposed YOLO based TP annotation method can accurately identify the targets with high annotation accuracies in complex backgrounds.

Deep learning-based object detection in low-altitude UAV datasets: A survey

As one of the hottest fields of computer vision, object detection is employed in many useful applications, including face recognition, autonomous cars and human behaviour analysis. CNN's have played a vital role in the progression of computer vision, especially object detection. In this paper, a review is provided on object detection methods which utilize deep learning to render outstanding results, and different object detection frameworks are discussed and analyzed with respect to their differences amongst each other. The comparisons are carried out based upon the criterion of accuracy and frames per second. The review starts with the explanation of CNNs and their advantages. Then we focus our attention on two types of object detection architectures and dive deep into their respective details. In the end, we draw a conclusion based on the facts detailed in the review.

Texture classification algorithms using local binary pattern (LBP) and its variants usually can achieve attractive results. However, the selected rotation invariant structural patterns in numerous LBP variants are not absolutely continuous invariant to any rotation angle. To improve the classification effectiveness on this occasion, in this paper, we introduce a robust descriptor based on the principal curvatures (PCs) and rotation invariant version of CLBP_Sign operator in completed LBP (CLBP), namely PC-LBP. Different from the original LBP and many LBP variants, PCs are employed in this paper to represent each local structure information due to their continuous rotation invariance. Simultaneously, both micro-and macro-structure texture information can also be captured through PCs, which comprise maximum and minimum curvatures. Inspired by the similar coding strategy of the CLBP_Sign operator, a new operator CLBP_PC is developed. By exploiting complementary information resulting from the two operators combination, the final PC-LBP descriptor has the properties of conspicuous rotation invariance, strong discriminativeness, gray scale invariance, needless of pretraining, and high computational efficiency. In addition, to improve the robustness of texture classification with multiresolution, a multiscale sampling approach is designed by adjusting three parameters accordingly. Experimental results demonstrate that the proposed multiresolution PC-LBP approach achieves comparable performance or outperforms a large number of state-of-the-art methods. Impressively, the classification accuracy of the proposed method performed on Outex_TC_00010 test suite is 100%.

Object detection based on deep learning has become one of the most widely used and popular research topics in the field of computer vision. This paper firstly reviews the development of target detection technology and analyzes and summarizes the advantages and disadvantages of traditional target detection methods. Then, several target detection technologies based on deep learning and their advantages and disadvantages are introduced in detail. Finally, the possible future development of object detection is discussed.

Despite significant advances in deep learning based object detection in recent years, training effective detectors in a small data regime remains an open challenge. This is very important since labelling training data for object detection is often very expensive and time-consuming. In this paper, we investigate the problem of few-shot object detection, where a detector has access to only limited amounts of annotated data. Based on the meta-learning principle, we propose a new meta-learning framework for object detection named which learns the ability to perform few-shot detection via meta-learning. Specifically, Meta-RCNN learns an object detector in an episodic learning paradigm on the (meta) training data. This learning scheme helps acquire a prior which enables Meta-RCNN to do few-shot detection on novel tasks. Built on top of the popular Faster RCNN detector, in Meta-RCNN, both the Region Proposal Network (RPN) and the object classification branch are meta-learned. The meta-trained RPN learns to provide class-specific proposals, while the object classifier learns to do few-shot classification. The novel loss objectives and learning strategy of Meta-RCNN can be trained in an end-to-end manner. We demonstrate the effectiveness of Meta-RCNN in few-shot detection on three datasets (Pascal-VOC, ImageNet-LOC and MSCOCO) with promising results.

SIFT (scale invariant feature transform) is one of the most robust and the widely used image matching algorithms based on local features. It has been used in the aspect of the cameras target matching. For its high complexity, this paper proposes a target matching method for multiple cameras by combining SIFT and the rotation-invariant LBP (local binary patterns). The object detection is obtained by improved Back-ground subtraction technique, and particle filter is used for region-based object tracking. The descriptor combining SIFT and the rotation-invariant LBP is calculated for every object region. The object matching is obtained by comparing the distance of feature matrix. Experimental results show that the proposed algorithm is more rapid than the standard SIFT algorithm while the performance is favorably compared to the standard SIFT algorithm, and it is robust in various scenes.

AeroNet: An efficient relative localization and object detection network for cooperative aerial-ground unmanned vehicles

To alleviate the costly data annotation problem in deep learning-based object detection, we leverage the canonical view model for active sample selection to improve the effectiveness of learning. Inspired by the view-approximation model, we hypothesize that visual features learned from canonical views denote better representations of objects, thus boosting the effectiveness of object learning. We validate the hypothesis empirically in the context of robot learning for novel object detection. Based on this, we propose a novel on-line viewpoint exploration (OLIVE) method that (1) defines goodness-of-view by combining informativeness of visual features and consistency of model-based object detection, and (2) systematically explores and selects viewpoints to boost learning efficiency. Furthermore, we train a legacy Faster R-CNN model with a data augmentation method while leveraging data samples generated by the OLIVE pipeline. We test our method on the T-LESS dataset and show that the proposed method outperforms competitive benchmarking methods, especially when the samples are few.

This paper presents an analysis of the statistical texture representation of the Local Binary Pattern (LBP) variants in the classification of wood defect images. The basic and variants of the LBP feature set that was constructed from a stage of feature extraction processes with the Basic LBP, Rotation Invariant LBP, Uniform LBP, and Rotation Invariant Uniform LBP. For significantly discriminating, the wood defect classes were further evaluated with the use of different classifiers. By comparing the results of the classification performances that had been conducted across the multiple wood species, the Uniform LBP was found to have demonstrated the highest accuracy level in the classification of the wood defects.

Objective To develop and evaluate an automated system for digitizing audiograms, classifying hearing loss levels, and comparing their performance with traditional methods and otolaryngologists' interpretations.Designed and Methods We conducted a retrospective diagnostic study using 1,959 audiogram images from patients aged 7 years and older at the Faculty of Medicine, Vajira Hospital, Navamindradhiraj University. We employed an object detection approach to digitize audiograms and developed multiple machine learning models to classify six hearing loss levels. The dataset was split into 70% training (1,407 images) and 30% testing (352 images) sets. We compared our model's performance with classifications based on manually extracted audiogram values and otolaryngologists' interpretations.Result Our object detection-based model achieved an F1-score of 94.72% in classifying hearing loss levels, comparable to the 96.43% F1-score obtained using manually extracted values. The Light Gradient Boosting Machine (LGBM) model is used as the classifier for the manually extracted data, which achieved top performance with 94.72% accuracy, 94.72% f1-score, 94.72 recall, and 94.72 precision. In object detection based model, The Random Forest Classifier (RFC) model showed the highest 96.43% accuracy in predicting hearing loss level, with a F1-score of 96.43%, recall of 96.43%, and precision of 96.45%.Conclusion Our proposed automated approach for audiogram digitization and hearing loss classification performs comparably to traditional methods and otolaryngologists' interpretations. This system can potentially assist otolaryngologists in providing more timely and effective treatment by quickly and accurately classifying hearing loss.