The vision-based fruit recognition and localization system is the basis for automatic operation of agricultural harvesting robots. Existing detection models are often constrained by high complexity slow inference speed, which do not meet real-time requirements Here, a method apple object proposed to address above problems. First, an improved YOLOX network designed detect target region, with multi-branch topology in training phase single-branch structure phase. spatial pyramid pooling layer...

ABSTRACT The detection of defects on industrial surfaces is essential for guaranteeing the quality and safety products. Deep learning‐based object methods have demonstrated impressive efficacy in applications recent years. However, due to complex variable shape defects, similarity between background, large intra‐class differences, small inter‐class differences lead low classification accuracy, it a great challenge achieve accurate defect detection. To overcome these challenges, this research...

The vision-based fruit recognition and localization system is the basis for operation of agricultural harvesting robots. However, there are many complicating factors in fruit-growing environment, such as changes light intensity, overlap between fruits, shading branches leaves, which increase difficulty robot vision localization. existing detection models often constrained by high complexity slow inference speed, do not meet real-time requirements Here, an apple-object method proposed to...

Intelligent apple-picking robots can significantly improve the efficiency of apple picking, and realization fast accurate recognition localization apples is prerequisite foundation for operation picking robots. Existing methods primarily focus on object detection semantic segmentation techniques. However, these often suffer from errors when facing occlusion overlapping issues. Furthermore, few instance are also inefficient heavily dependent results. Therefore, this paper proposes an method...

<abstract> <p>As an essential part of electronic component assembly, it is crucial to rapidly and accurately detect components. Therefore, a lightweight detection method based on knowledge distillation proposed in this study. First, student model was constructed. Then, we consider issues like the teacher student's differing expressions. A combination feature channel learn teacher's rich class-related inter-class difference features. Finally, comparative experiments were analyzed...

Achieving multi-scene electronic component detection is the key to automatic assembly. The study of a deep-learning-based object method an important research focus. There are many anchors in current methods, which often leads extremely unbalanced positive and negative samples during training requires manual adjustment thresholds divide samples. Besides, existing methods bring complex model with parameters large computation complexity. To meet these issues, new was proposed for components...

The detection of defects on industrial surfaces is essential for guaranteeing the quality and safety products. Deep learning-based object methods have demonstrated impressive efficacy in applications recent years. However, surface are complex variable shape, similar to background features, inter-class feature differences minor, while intra-class large, making it a significant challenge achieve accurate different defect categories. To overcome these challenges, this research proposed novel...

Tea bud detection and picking point localization in natural environments are essential issues robotic harvesting. However, the high growth density varying attitude of tea buds bring excellent challenges to model. Therefore, this study proposes a computer vision system combining deep learning machine techniques accurately acquire guide robot harvest buds. Firstly, an improved lightweight Mask R-CNN model is proposed detect segment The adopts MobileViT as backbone network for feature...

The continuous development of artificial intelligence technology has promoted the construction smart libraries and their intelligent services. In process access to books, extraction requested book number region become an important part process. is generally affixed bottom spine book, which small in size, height not always same, so it’s difficult identify. By way, due images’ resolution, shooting angle other practical problems, difficulty work will be increased. To improve identification...