激光与光电子学进展, 2020, 57 (5): 053004, 网络出版: 2020-03-05
基于高光谱成像技术的不同介质血迹陈旧度研究 下载: 933次
Age Estimation of the Bloodstains on Different Substrates Based on the Hyperspectral Imaging Technology
光谱学 高光谱成像 血迹 特征光谱 人工神经网络 遗留时间 spectroscopy hyperspectral imaging bloodstain characteristic spectrum artificial neural networks aging time
摘要
提出一种基于高光谱成像技术的血迹陈旧度无损预测方法。利用高光谱成像仪获取白瓷砖、白纸、白棉布3种介质上不同陈旧度的血迹光谱图像,筛选550~800 nm特征光谱波段,并通过人工神经网络模型实现血迹遗留时间的预测。结果表明:在实验室环境下,30 h内3种介质上的血迹遗留时间预测的回归分析决定系数( RP2)均大于0.9930,预测方均根误差(RMSEP)为18.29 min,中位相对误差(δM)平均值为8.05%,可见该方法能对较短时间内的血迹陈旧度进行有效预测。
Abstract
In this study, we propose a nondestructive method based on hyperspectral imaging technology to determine the bloodstain age. The spectral images of the bloodstains on a white ceramic tile, white paper, and white cotton cloth with different aging times can be obtained using a hyperspectral imager and can be characterized in the spectral range of 550-800 nm. Further, the bloodstain age is predicted using an artificial neural network model. The results indicate that the regression coefficients of the aging time of bloodstains for the three samples are larger than 0.9930 within 30 h under the laboratory conditions. The average root mean square error of prediction (RMSEP) is 18.29 min, and the median relative error δM is 8.05%. Thus, the proposed method is observed to be effective for estimating the aging time of the bloodstains in a short time (<30 h).
冯颖, 蔡竞. 基于高光谱成像技术的不同介质血迹陈旧度研究[J]. 激光与光电子学进展, 2020, 57(5): 053004. Ying Feng, Jing Cai. Age Estimation of the Bloodstains on Different Substrates Based on the Hyperspectral Imaging Technology[J]. Laser & Optoelectronics Progress, 2020, 57(5): 053004.