Mu Hua Liu

Preliminary Study on Quantification of Duck Color Based on Fuzzy K – Nearest Neighbor Method

In order to quantify duck color quickly and accurately, a visible light image acquisition system was developed in this study. Firstly, duck breast meat was extraced from the original image using over-threshold segmentation method; and then image fleshcolor features including average value of L*，a* and b* were extracted through color space conversion; finally, fleshcolor features were used as characteristic parameters to establish quantification model for duck color based on the method of Fuzzy k-nearest neighbor (F KNN) and BP neural network and the experimental datas gotten from the two methods were analyzd in contrast. The results showed the accuracy rate of test samples by F KNN method was higher than that of BP neural network, the accuracy rate was 83.78% and 78.38% respectively. Therefore, F KNN method was selected for quantification of duck color in this research.

Classification Modeling of Parts for Complex Machinery Product Based on Design Structure Matrix

In order to reuse existing resources effectively and meet the specific demands of different users, it is necessary to study the classification modeling methods of parts for guiding product design and improving product quality. Firstly, the hierarchy theories of classification modeling are analyzed based on Design Structure Matrix (DSM). Then, the process of classification modeling is presented for parts of complex machinery product based on DSM modeling method. Meanwhile, the steps and notes to build classification structure were described in details. The results show that the theories and methods presented in this article could provid guidances for classification modeling of parts and realize efficient reuse and quick retrieval of product data for complex machinery manufactories.

Study of Pesticide Contaminated Navel Orange Recognition Using near Infrared Spectroscopy

Based on Support Vector Machine (SVM) and genetic algorithm (GA), this paper intends to search for the characteristic spectral ranges and wavelengths of near infrared spectroscopy of navel oranges contaminated by different pesticides, and set up recognition models. The pesticides in the experiment were Lannate®L insecticide, fenvalerate and omethoate, and three different concentrations were given to each pesticide. Preparing ten groups of navel oranges, each group was sprayed with a different pesticide and the 10th group without pesticide spraying was used for comparison. Searching the whole spectral range through GA, 5 best spectral ranges (165 wavelengths) were obtained and the recognition rate reached 98.86%. Then based on the chosen spectral ranges, 85 feature wavelengths were extracted with continual GA-SVM optimization, and the recognition rate was 99.14%. Experiment results showed that the application of SVM combining with GA can not only improve recognition accuracy, but also simplify the model effectively

Design of Poultry Meat Edible Quality Detection Device Using Multiple Image Sensors

A detection device was designed to detect the poultry meat edible quality. The detection device is the electromechanical integration equipment which is integrated with x-ray technology, visible light technology and induced fluorescence technology. The device not only can overcome the disadvantages that the conventional measure methods have destructiveness to the samples, and consume the long time of the detection, etc, but also can detect the multiple quality indexes simultaneously. The detection device is composed of the frames, the transmission control system, the X-ray detection system, the visible light detection system and the induced fluorescence detection system, etc. The device can greatly increase the judging ability for comprehensive quality of poultry meat and realize the fast and non-destructive testing to poultry meat edible quality.

Application of Three-Dimensional Fluorescence Spectroscopy Coupled with ATLD in Rapid Determination of Triazophos Content in Duck Meat

The triazophos is a kind of organic phosphorus pesticide, the quantitative determination model based on three-dimensional fluorescence spectroscopy coupled with alternating trilinear decomposition (ATLD) was explored for realizing the rapid determination of triazophos content in duck meat. Firstly, three-dimensional fluorescence spectra of duck meat extract, triazophos standard solution and duck meat extract containing triazophos were explored, respectively. Secondly, the fluorescence quenching phenomenon of the triazophos in duck meat extract was analyzed. Lastly, the number of components for three linear decomposition of ATLD was set as 2 by using the core consistency diagnostic, and the calibration curve between the relative fluorescence intensity and the actual concentration of the triazophos was established by using ATLD. The experimental results showed that the determination coefficient (R2) and the root mean squared error of prediction (RMSEP) for the proposed model in this paper were 0.9741 and 0.764 respectively, and it was feasible to predict the triazophos content in duck meat combining with three-demensional fluorescence spectroscopy and ATLD.

Determination of Moisture Content in Ginger Using PSO Combined with Vis/NIR

Vis/NIR spectroscopy was used to measure the moisture content of ginger. 330 samples were separated into two groups, as training and validation. Vis/NIR reflection spectral data from 350 to 1800 nm were collected using ginger within the training and validation sets. PSO was used to establish the PLS model. In comparison to the full spectrum model (contained 1451 variables), the prediction capability was improved after using PSO for PLS models. The number of selected variables and LVs were 300 and 6, respectively. The correlation of determination in validation set (), root mean square error of prediction (RMSEP), and bias by PSO-PLS were 0.9881, 4.7827, and 0.1751.

Nondestructive Detection of Soluble Solids Content of Nanfeng Mandarin Orange Using VIS-NIR Spectroscopy

This study demonstrated how VIS-NIR spectroscopy can be used in the quantitative, noninvasive probing of soluble solids content (SSC) of mandarin orange. Total 197 mandarin oranges were divided into calibration set (133 samples) and prediction set (64 samples). Multiple scatter correction (MSC) was used to preprocess the collected visible and near infrared (Vis-NIR) spectra (350-1800nm) of mandarin orange. Partial least square (PLS), interval partial least square (IPLS) and synergy interval partial least square (SIPLS) methods were applied for constructing predictive models of SSC. Experimental results showed that the optimal SIPLS model obtained with 10 PLS components and the optimal combinations of intervals were number 5,7,8,9. The correlation coefficient (r) between the predicted and actual SSC was 0.9265 and 0.8577 for calibration and prediction set, respectively. The root mean square error of calibration (RMSEC) and prediction (RMSEP) set was 0.4890 and 0.7113, respectively. In conclusion, the combination of Vis-NIR spectroscopy and SIPLS methods can be used to provide a technique of noninvasive, convenient and rapid analysis for SSC in fruit.

Study of Fluorescence Spectrum for Measurement of Soluble Solids Content in Navel Orange

On-line nondestructive estimation of fruit internal quality greatly benefits the consumer and the fruit industry as a whole because it can ensure the safety and processed product consistency. This research was aimed to build prediction model of soluble solids content (SSC) in navel oranges using ultraviolet lamp induced fluorescence technique. Total 254 navel oranges were separated into the calibration set (170 samples) and prediction set (84 samples). The actual SSC of navel oranges were measured using a digital refratometer. A characteristic wavelength range (385-414nm) was found to build a partial least squares (PLS) model for predicting the SSC in navel oranges. The correlation coefficient (r) between the predicted and actual SSC was 0.9161 and 0.8295 for calibration and prediction set, respectively. The root mean square error of calibration (RMSEC) and prediction (RMSEP) set was 1.0797 and 1.4283, respectively. The results showed that fluorescence spectrum can nondestructively determine SSC in fruit.

Detection of Defect on Navel Orange Using Hyperspectral Reflectance Image

A hyperspectral imaging system for detecting defect on navel orange was demonstrated. The hyperspectral imaging system, which was a line-scan imaging system, consisted of a hyperspectral camera, a halogen lighting unit, a computer and a translation stage. The imaging system operated from 400 to 1000nm. Principal component analysis (PCA) was performed using the hyperspectral images data (from 500 to 700nm); 2nd principal component (PC) image exhibited differential responses between normal and defect spots on the surface of navel orange. The combined use of the PC-2 images demonstrated the detection of defect spots with minimal false positives. Based on the PC-2 weighing coefficients, the dominant wavelengths were 528,529,530,673,674 and 675nm. This research demonstrated the potential of multispectral image for online applications for detection of defect on navel oranges.

Study on Mass and Thickness Measurement of Chicken Meat Using X-Ray Image

The mass and thickness measurement of chicken meat were investigated using X-ray image in the paper. The relation model between pixel value of X-ray image and thickness of chicken breast was established using the exponential fitting and the exponential fitting equation is y = 2610.2e-0.0283x. The relation model between pixel value of X-ray image and average mass of chicken breast was established using the exponential fitting and the exponential fitting equation is y =0.3575e-0.0308x. The experimental results showed that the modeling regression coefficient and predicted regression coefficient for thickness measurement were 0.9604 and 0.976, respectively. The modeling regression coefficient and predicted regression coefficient for mass measurement were 0.844 and 0.881, respectively.