基于激光传感信号的动态环境目标智能识别研究

Automatic recognition of moving objects in dynamic environment based on machine vision

由于利用原有方法自动识别复杂环境运动目标时,图像存在的噪声会导致误识别率较高,因此,提出一种基于机器视觉的复杂环境运动目标自动识别方法。通过高斯混合建模方法检测复杂环境下的运动目标,具体步骤为确定高斯混合模型的实际初始参数,更新高斯混合模型的对应参数,最后对各高斯函数的实际权值实施归一化处理。基于机器视觉中的局部二值模型提取复杂环境下的运动目标特征,根据提取的运动目标特征,通过三帧差法选取并比较三帧相邻的复杂环境图像,然后两两作差并实施二值化,以及执行逻辑“与”运算,实现复杂环境下运动目标特征的识别。通过将基于机器视觉的复杂环境运动目标自动识别方法与原有方法进行性能对比实验,证明该方法的误识别率低于两种原有方法,实现了运动目标识别性能的提升,对于监控系统的发展有很大意义。

When the original method is used to automatically recognize moving targets in a complex environment,the noise in the image will lead to a high false recognition rate.Therefore,a method of automatic recognition of moving objects in a dynamic environment based on machine vision is proposed.The specific steps are to determine the actual initial parameters of the Gaussian mixture model,update the corresponding parameters of the Gaussian mixture model,and finally normalize the actual weights of each Gaussian function.Based on the local binary model of machine vision,the feature of moving object in complex environment is extracted.According to the feature of moving object,three adjacent complex environment images are selected and compared by three frame difference method,and then the difference is made by two and binarization is implemented,and the logical"and"operation is performed to realize the recognition of moving target features in dynamic environment.By comparing the performance of the automatic recognition method based on machine vision and the original method in dynamic environment,it is proved that the error recognition rate of the method is lower than that of the two original methods,and the performance of the moving target recognition is improved,which has great significance for the development of the monitoring system.