Search

Object detection in Unmanned Aerial Vehicle (UAVs) faces a significant challenge in computer vision. Traditional methods are difficult to model object appearance feature with large scale variations and viewpoint differences, when drones fly at different altitudes and capture images from diverse shooting angles. To address this issue, we propose a Learnable Cross-scale Sparse Attention (LCSA) guided feature fusion method to improve the performance of UAV object detection. Specifically, the LCSA feature fusion module enables each point in a feature map to aggregate discriminative information from a set of points with learnable offsets in neighbor feature maps. It enhances local discriminative features of the object by facilitating semantic information interaction across multiple feature maps. The LCSA can function as a novel neck method that complements the existing neck methods and is also transplantable to different object detection frameworks. Moreover, we also employ a scale-aware loss function to integrate the normalized Wasserstein distance with CIoU in order to improve the incompatibility of IoU for objects with large scale variance. Experimental results on the SeaDroneSeev2 and VisDrone2019-DET datasets show that the proposed method achieves superior performance. At a resolution of 640*640, our method achieves 81.9% AP50 and 47.4% AP on SeaDroneSeev2, surpassing baseline 4.9% and 4.8%, achieves state-of-the-art performance. Furthermore, our method outperforms baseline by 5% AP on VisDrone2019-DET. Code will be available at https://github.com/qch777/LSACF.

Abstract High quality feature fusion module is one of the key components for multispectral pedestrian detection system in challenging situations, such as large‐scale variance and occlusion. Although attention mechanism is one of the most effective ways for feature refining, the correlation between attention and scales in feature pyramid still remains unknown. Therefore, a scale‐aware permutated attention module is proposed to enhance features of objects with different scales adaptively in the feature pyramid. Specifically, four different local and global attention sub‐modules are investigated to refine feature maps with different permutations in the Feature Pyramid Networks, improving the quality of the feature fusion. Besides, to address the high miss‐rate issue for small‐sized pedestrians, an adjacent‐branch feature aggregation module is proposed to aggregate features across different scales, taking both semantic context and spatial resolution into consideration. Both modules can benefit from each other with significant performance improvement in terms of efficiency and accuracy, when equipped with the dual‐branch CenterNet detection framework. Experiments on the KAIST and FLIR datasets demonstrate its superior performance compared with other state‐of‐the‐arts.

Multispectral pedestrian detection system suffers with low feature fusion quality, high quantity of model hyper-parameters and complex anchor matching algorithm.To deal with these problems, an anchor free multispectral pedestrian detection algorithm based on differential feature attention mechanism was proposed.Firstly, differential modality aware fusion was used to obtain the complementary information between different modalities to optimize the channel features.Secondly, the CenterNet detection framework with anchor free mechanism was adopted to greatly reduce the computational complexity of the model and thus improve the detection speed.Finally, differential feature guided attention mechanism was introduced to improve the quality of feature fusion and further enhance the detection accuracy.Experimental results on three open datasets, KAIST, CVC14 and FLIR, show that the proposed algorithm can effectively improve the detection accuracy and speed compared with the current advanced methods, and has a good practical application prospect.

Apple diseases cause a lot of economic losses to fruit growers in China. Early diagnosis and accurate recognition of apple diseases can control the spread of disease and reduce production costs. However, the significance of disease characteristic of apple leaves in complex environment is relatively weak, and the fine-grain among different diseases of apple leaves is high, and the conventional feature extraction methods will lose the discrimination information. To solve these problems, an apple disease classification model based on multi-scale feature fusion is proposed in this paper. Firstly, the information flow of conventional residual network (ResNet) was improved to achieve efficient information circulation through changing the position of batch normalization and rectified linear unit (ReLU). Secondly, in order to solve the problem of serious loss of information in ResNet downsample, the channel projection and spatial projection of downsample were separated. Lastly, the $3\times 3$ conv in ResBlocks was replaced by pyramid convolution, and the dilated convolution with different dilation rate was introduced into pyramid convolution to enhance the output scale of feature maps and improve the robustness of the model. The optimized model was verified on the dataset of this paper, and the optimized model had stronger anti-noise ability and better robustness, excellent learning effect and fast convergence speed. The classification accuracy on the original dataset is 94.24%, and that on the preprocessed dataset is 94.99%. The results demonstrate that the optimal model has a high accuracy, which can provide a reference for the prevention and control of apple leaf diseases.

Conventional AlexNet has the problems of slow training speed, single characteristic scale and low recognition accuracy. To solve these problems, a convolutional neural network identification model based on Inception module and dilated convolution is proposed in this study. The inception module combined with dilated convolution, could extract disease characteristics at different scales and increase the receptive field. By setting different parameters, six improved models were obtained. They were trained to identify 26 diseases of 14 different crops; then the authors selected optimal recognition model. On this basis, the segmented dataset and the grey‐scaled dataset were trained as comparative experiments to explore the influence of background and colour features on the recognition results. After only two training epochs, the improved optimal model could achieve an accuracy of over 95%. Moreover, the final average identification accuracy reached 99.37%. Contrast experiments indicate that colour and background features may influence the recognition effect. The improved model can extract disease information from different scales in the feature map to identify diverse diseases of different crops. The proposed model has faster training speed and higher recognition accuracy than the traditional model, and thus it can provide a reference for crop disease identification in actual production.

Combining disease categories and crop species leads to complex intra-class and inter-class differences. Significant intra-class difference and subtle inter-class difference pose a great challenge to high-precision crop disease classification tasks. To this end, we propose a multi-granularity feature aggregation method for accurately identifying disease types and crop species as well as better understanding the disease-affected regions implicitly. Specifically, in order to capture fine-grained discriminating clues to disease categories, we first explored the pixel-level spatial self-attention to model the pair-wise semantic relations. Second, we utilized the block-level channel self-attention to enhance the feature-discriminative ability of different crop species. Finally, we used a spatial reasoning module to model the spatial geometric relationship of the image patches sequentially, such that the feature-discriminative ability of characterizing both diseases and species is further improved. The proposed model was verified on the PDR2018 dataset, the FGVC8 dataset, and the non-lab dataset PlantDoc. Experimental results demonstrated our method reported respective classification accuracies of 88.32%, 89.95%, and 89.75% along with F1-scores of 88.20%, 89.24%, and 89.13% on three datasets. More importantly, the proposed architecture not only improved the classification accuracy but also promised model efficiency with low complexity, which is beneficial for precision agricultural applications.

The yield and security of grain are seriously infringed on by crop diseases, which are the critical factor hindering the green and high-quality development of agriculture. The existing crop disease identification models make it difficult to focus on the disease spot area. Additionally, crops with similar disease characteristics are easily misidentified. To address the above problems, this paper proposed an accurate and efficient disease identification model, which not only incorporated local and global features of images for feature analysis, but also improved the separability between similar diseases. First, Transformer Encoder was introduced into the improved model as a convolution operation, so as to establish the dependency between long-distance features and extract the global features of the disease images. Then, Centerloss was introduced as a penalty term to optimize the common cross-entropy loss, so as to expand the inter-class difference of crop disease characteristics and narrow their intra-class gap. Finally, according to the characteristics of the datasets, a more appropriate evaluation index was used to carry out experiments on different datasets. The identification accuracy of 99.62% was obtained on Plant Village, and the balanced accuracy of 96.58% was obtained on Dataset1 with a complex background. It showed good generalization ability when facing disease images from different sources. The improved model also balanced the contradiction between identification accuracy and parameter quantity. Compared with pure CNN and Transformer models, the leaf disease identification model proposed in this paper not only focuses more on the disease regions of leaves, but also better distinguishes different diseases with similar characteristics.

In this paper, a shallow–deep feature fusion (SDFF) method is developed for pedestrian detection. Firstly, we propose a shallow feature-based method under the ACF framework of pedestrian detection. More precisely, improved Haar-like templates with Local FDA learning are used to filter the channel maps of ACF such that these Haar-like features are able to improve the discriminative power and therefore enhance the detection performance. The proposed shallow feature is also referred to as weighted subset-haar-like feature. It is efficient in pedestrian detection with a high recall rate and precise localization. Secondly, the proposed shallow feature-based detection method operates as a region proposal. A classifier equipped with ResNet is then used to refine the region proposals to judge whether each region contains a pedestrian or not. The extensive experiments evaluated on INRIA, Caltech, and TUD-Brussel datasets show that SDFF is an effective and efficient method for pedestrian detection.

In the current natural environment, due to the complexity of the background and the high similarity of the color between immature green tomatoes and the plant, the occlusion of the key organs (flower and fruit) by the leaves and stems will lead to low recognition rates and poor generalizations of the detection model. Therefore, an improved tomato organ detection method based on convolutional neural network (CNN) has been proposed in this paper. Based on the original Faster R-CNN algorithm, Resnet-50 with residual blocks was used to replace the traditional vgg16 feature extraction network, and a K-means clustering method was used to adjust more appropriate anchor sizes than manual setting, to improve detection accuracy. The test results showed that the mean average precision (mAP) was significantly improved compared with the traditional Faster R-CNN model. The training model can be transplanted to the embedded system, which lays a theoretical foundation for the development of a precise targeting pesticide application system and an automatic picking device.