【摘要】 为解决当前基于遥感图像识别技术中的错误判断问题及遗漏对象问题等缺陷,提出一种改进的YOLOv8目标检测算法来提升其准确率。使用Efficient ViT作为主要框架替代原始架构设计;在核心部分嵌入深度分离式卷积单元,构建Slim Neck构架结合CA注意力机制和Ghostmodules;提出Focal-IOU损失函数重构IOU,取代CIOU。实验结果表明,改进后算法与原YOLOv8算法相比,m AP提升了2.4%,精确度和召回率也有一定提升,验证了改进算法的有效性和先进性。更多还原

【Abstract】 To solve the shortcomings of the current remote sensing image recognition technology,such as the problem of misjudgment and the problem of missing objects,an improved YOLOv8 object detection algorithm was proposed to improve its accuracy.The EfficientViT was used as the primary framework to replace the original architecture design.In the core part,a deep separated convolutional unit was embedded,and the SlimNeck architecture was constructed combining the CA attention mechanism and Ghostmodulcs.The Focal-IOU loss function was proposed to reconstruct the IOU to replace the CIOU.Experimental results show that compared with the original YOLOv8 algorithm,the mAP of the improved algorithm is increased by 2.4%,and the accuracy and recall rate are also improved to a certain extent,verifying the effectiveness and advancement of the improved algorithm.更多还原