Grasping a specified object from multi-object scenes is an essential ability for intelligent *** ability depends on the affiliation between the grasp position and the object category. Most existing multi-object grasp ...
详细信息
Grasping a specified object from multi-object scenes is an essential ability for intelligent *** ability depends on the affiliation between the grasp position and the object category. Most existing multi-object grasp detection methods considering the affiliation rely on object detection results, thus limiting the improvement of robotic grasp detection accuracy. This paper proposes a decoupled single-stage multitask robotic grasp detection method based on the Faster R-CNN framework for multi-object scenes. The designed network independently detects the category of an object and its possible grasp positions by using one loss function. A new grasp matching strategy is designed to determine the relationship between object categories and predicted grasp positions. The VMRD grasp dataset is used to test the performance of the proposed method. Compared with other grasp detection methods, the proposed method achieves higher object detection accuracy and grasp detection accuracy.
This paper presents a voice-assisted pick and place system using a mobile manipulator. The system consists of a voice recognition module, a Kinova Gen 3 Lite arm, and a Jackal mobile base. The voice recognition module...
详细信息
Insulator inspection is important for high-voltage transmission line safety. However, existing object-detection algorithms cannot complete the task robustly in complex aerial environments. To improve the performance o...
详细信息
Insulator inspection is important for high-voltage transmission line safety. However, existing object-detection algorithms cannot complete the task robustly in complex aerial environments. To improve the performance of the detection task, this paper proposes an algorithm, called Faster R-Transformer, which combines the advantage of Convolutional neural networks and the self-attention mechanism for aerial insulator detection. First, the D-ResNet performs feature extraction to obtain a larger receptive field. Then, the RPN module is used for region recommendation, and the Align Pooling module is used for region Pooling. Finally, the self -attention mechanism is adopted in the proposed region to perform local region feature weighting. The Faster R-Transformer was evaluated in complex aerial environments, and the evaluation showed that the proposed method was capable of detecting insulators of various sizes under different illuminations and viewpoints. We also explored hyperparameters of Faster R-Transformer to meet the requirements of various scenarios.
Object detection is to locate and recognize the object in the ***,the existing object detection methods are prone to miss detection or false detection on small objects because small objects have few pixels(less inform...
详细信息
Object detection is to locate and recognize the object in the ***,the existing object detection methods are prone to miss detection or false detection on small objects because small objects have few pixels(less information)and more significant interference from the background. This thesis aims to reduce the miss detection rate of small objects and improve detection accuracy based on the existing object detection *** thesis proposes Scale Adaptive Feature Pyramid Networks(SAFPN)*** has employed weights chosen adaptively to each input image in fusing feature maps of the bottom-up pyramid and top-down *** adaptive weights are computed by using a scale attention module built into the feature map fusion *** thesis also proposes and evaluates another novel object detection architecture called Cascaded Multi-Channel Feature Pyramid Network(CMFPN).The CMFPN''s architecture employs cascaded feature pyramids to obtain salient,a highly semantic feature pyramid for object location proposal regression and *** architecture uses deeper channels for the more semantic stages by using the cascade so that the salient features are extracted and preserved. In this thesis,comparative experiments are carried out on open datasets(Pascal VOC datasets).The experimental evaluation demonstrated the proposed approaches''effectiveness.
The outdoors insulator is an important component of the high-speed railway traction substation,which is of great significance to maintain the stability of transmission line and ensure the normal operation of transmiss...
详细信息
The outdoors insulator is an important component of the high-speed railway traction substation,which is of great significance to maintain the stability of transmission line and ensure the normal operation of transmission *** there is a fault for the insulator,it will cause serious transmission failure and economic ***,a method is proposed to detect the abnormal areas of outdoors insulator in high-speed railway traction substation based on object detection and generative adversarial ***,we employ Faster RCNN to locate the area of insulator from the input image of traction ***,the image of insulator obtained from the first step is fed into our designed generative adversarial networks to generate fake image,which is a normal image of ***,multi-scale structural similarity algorithm is used to realize the anomaly detection of insulator and visualize anomalous *** results on Heishan traction substation show that the proposed method is effective.
Human attention is an important cognitive resource to take into account when designing effective human-machine interaction and cognitive computing systems. Much of our knowledge about attention processing stems from s...
详细信息
ISBN:
(纸本)9781479960811
Human attention is an important cognitive resource to take into account when designing effective human-machine interaction and cognitive computing systems. Much of our knowledge about attention processing stems from search tasks that are usually framed around Treisman's feature integration theory. However, search performance in these tasks has mainly been investigated using an overt attention paradigm. Covert attention on the other hand has hardly been investigated in this context. To gain a more thorough understanding of human attentional processing, we have experimentally compared search performance when people are instructed to either overtly or covertly search for targets under a variety of target/distractor combinations. The overt search results presented in this work agree well with the guided search studies by Wolfe et al. While response times are similar between the two attention conditions, we found that error rates are considerably higher in covert search.
暂无评论