【摘要】 针对圈养猪夜间难以监测的问题,提出一种基于改进SSD的夜间猪群舒适度监测方法。首先,在SSD(single shotmultiboxdetector)算法中引入焦点损失函数,解决夜间目标与背景差异小而难以识别的问题;其次,使用soft-NMS算法检测遮挡的猪只目标;再次,通过二次迁移学习方法得到夜间圈养猪目标检测模型;最后,分析猪只位置信息监测夜间猪群舒适度。结果表明,所提方法能精确检测夜间多只猪目标,并对其舒适度进行监测。更多还原

【Abstract】 Aiming at the problem that it is difficult to monitor pigs at night,a method for monitoring the comfort of pigs at night based on improved SSD is proposed. First,the focus loss function is introduced in the SSD(single shot multibox detector)algorithm to solve the problem that the difference between the pig target and the background at night is difficult to identify. Secondly,the soft-NMS algorithm is used to detect occlusion pig targets. Thirdly,the night pig target detection model is obtained through the secondary transfer learning method. Finally,analyze the location information of pigs to monitor their comfort at night. The results show that the proposed method can accurately detect multiple pig targets at night and monitor their comfort.更多还原