Foreign Object Detection And Grade Discrimination Of Sausages Based On Ultrasonic Imaging

Sausage is a popular meat product with nutrient-rich and easy to store procedures.However,foreign objects such as glass,plastic,mental can enter into sausage through the machining process and adulterated sausage still happened in the market.These quality problems have greatly impaired the legitimate interests of consumers.Therefore,it is important to detect foreign objects in sausage and it is also important to identification the grades of sausages In this study,the common foreign objects in sausage were detected and identified based on ultrasonic imaging technology,along with sausage grades recognition were carried out at the same time,the main research work and conclusion are summarized as follows:(1)Optimization of ultrasonic imaging parameters and calibration of sausage grades.Through the analysis of each ultrasonic imaging parameter and the imaging results of excellent grades sausage of brand A,ultrasonic imaging parameters were optimized.The optimized parameters were as shown below: the pulse voltage;300 V,the pulse width;25 ns,the pulse repetition frequency;800 Hz,the gain;35 dB,the resolution;0.1 mm,the focal length;1″,and the scanning rate was 5 mm/s.The principal composition of sausage was determined according to the national standard of China(GB 5009.5-2016,GB 5009.9-2016,GB 5009.3-2016,GB 5009.6-2016,);thereafter,grades of sausages collected from brand A and B were calibrated based on the results obtained.Results showed that there were significant differences(P<0.05)among the analyzed contents of each grade of sausages.Premium quality sausages had best internal integrity and stability with the highest protein and lowest starch content.Whereas,general sausage indicated floppy interior,poor overall quality with the lowest protein content and highest starch content.The general sausage also exhibited a certain granularity.(2)Foreign object detection and recognition based on ultrasonic imaging technology.Different sizes(3×3 mm,5×5 mm and 10×10 mm)of metal,glass,and plastic were embedded in different depths(5 mm,10 mm,15 mm)of the luncheon sausages,and echo signals and ultrasound image acquisition were performed.Luncheon sausage without foreign object was used as reference group.It was observed that the echo signals(magnitude,period,etc.)of each foreign object were significantly different from the reference group.Detection depth of foreign objects was approximately 15 mm at 20 MHz transducer.The automatic threshold segmentation was used to analyze the ultrasonic images of luncheon sausage with foreign objects.The foreign objects detection rate for all samples was 96.29%.The foreign object identification model was established with the average echo intensity,area detection rate,and common texture feature(angular second moment,contrast and so on)using support vector machine(SVM)model,when the number of principal components(PCs)6,and the calibration and prediction set recognition rate were 83.27% and 81.67%,respectively.(3)Grade identification of sausage based on ultrasonic imaging technology.With the sausages calibrated by(1),all the samples were used for texture measurement and ultrasonic image acquisition.The results of texture analysis showed that there was little difference in the texture values of the same grade sausages under the same brand,but the texture values of different grade sausages were significantly difference(P<0.05).The inside of the premium sausage had the best binding strength,integrity and stability,while the internal parts of the general sausage had weak binding force,worst integrity and stability,and a certain granularity.The results of ultrasonic imaging showed that there were obvious differences among different grades,but the same grade with different brand were similar.Textural features extracted from different grade images had significant difference(P<0.05).Trend of textural feature values were in accordance with ultrasound images.All the samples of sausage were divided into three classes by principal component analysis(PCA)according to the grade of the sample.From the grades determination model established with texture features,the support vector machine(SVM)model had the best recognition rate,with rates of calibration set and prediction sets up to 100%.In this study,foreign objects detection and grades rapid identification of sausages were realized by ultrasonic imaging technology combined with image processing and pattern recognition technology.The research results provided a reference for foreign objects detection and grades rapid identification of sausages by ultrasonic imaging technology.And it is beneficial to the application of ultrasonic imaging technology in the internal quality inspection of food.