CAJ | 학술논문

为了有效提高野外作业的监管效率、保障目标区域安全、满足监控智能化需求，提出了一种基于数字信号处理器(DSP)的野外作业车辆智能监控系统设计方案。 DSP平台、太阳能供电、3G无线数据传输三大模块的设计构成了野外作业监控系统的硬件部分。软件部分提出了一种基于梯度信息的视觉背景提取( ViBe)方法，可以有效滤除目标的阴影部分，提高检测的鲁棒性。提出了一种基于前景分割的自适应跟踪算法，保证目标被准确跟踪。实验表明，系统能够每天24小时持续稳定地对车辆进行自动检测、跟踪和统计，准确率达到96%以上，并且具有在其他领域的迁移应用能力。
위료유효제고야외작업적감관효솔、보장목표구역안전、만족감공지능화수구，제출료일충기우수자신호처리기(DSP)적야외작업차량지능감공계통설계방안。 DSP평태、태양능공전、3G무선수거전수삼대모괴적설계구성료야외작업감공계통적경건부분。연건부분제출료일충기우제도신식적시각배경제취( ViBe)방법，가이유효려제목표적음영부분，제고검측적로봉성。제출료일충기우전경분할적자괄응근종산법，보증목표피준학근종。실험표명，계통능구매천24소시지속은정지대차량진행자동검측、근종화통계，준학솔체도96%이상，병차구유재기타영역적천이응용능력。
In order to effectively improve the efficiency of supervision, ensure the safety of operations and meet the need of intelligent monitoring in the outdoor operation, a design scheme of vehicle monitoring system for field operation was presented. DSP platform, solar energy and 3G wireless data transmission was designed for using in outdoor surveillance, providing a hardware foundation. In the part of software, an improved ViBe ( Visual Background Extractor) modeling algorithm and a new adaptive tracking algorithm based on foreground segmentation were proposed, which can effectively erase the shadows and accurately track vehicles in field environment. Experiments show that the system can steadily, real-time run 24 hours a day with accuracy of above 96%, and can be used in other fields.