猪肉颜色是猪肉质量中消费者最为直观，且最具重要意义的一个指标，直接决定消费者的购买欲望，它受品种、饲养方式、季节环境、宰前操作等因素影响。目前我国缺乏猪肉颜色和加工特性相接合的分级技术，屠宰企业不根据肉的颜色进行分级，加工企业不能根据肉的加工特性进行合理加工，给企业带来不小的经济损失。本论文基于感官评分、仪器法和机器视觉技术，研究了不同颜色等级猪肉的加工特性，通过找出颜色变量的加工特性预测指标，初步建立仪器法和机器视觉技术预测方法，并对肉糜加工特性的预测方法进行探索，为生猪养殖业、屠宰企业、肉类深加工企业和消费者提供技术理论指导。 第一章：以180头猪胴体背最长肌为研究对象，对猪肉色泽进行感官评定，研究了1-6颜色等级猪肉匀浆物的蛋白溶解度、凝胶强度和弹性、凝胶保水性等加工特性之间的关系。在离子强度0.45M、0.60M NaCl和pH 6.5条件下制备1-6颜色等级猪肉匀浆物，测定了其蛋白溶解度、凝胶强度和弹性、凝胶保水性等加工特性。结果表明：感官评分和离子强度对猪肉匀浆物的蛋白溶解度、凝胶强度、凝胶弹性和凝胶保水性均有极显著性影响（P<0.01）。高离子强度组的凝胶强度、凝胶弹性、凝胶保水性和蛋白溶解度均显著大于低离子强度组的（P<0.05）。感官评分和离子强度对猪肉凝胶强度有互作效应（P<0.05），对蛋白溶解度、凝胶弹性和凝胶保水性均无互作效应（P>0.05）；感官评分、离子强度与蛋白溶解度、凝胶强度、凝胶弹性和凝胶保水性之间均呈极显著正相关（P<0.05）。 第二章：基于机器视觉技术研究了加工特性与猪肉H、S、L颜色变量之间的关系。在感官评分之后，利用计算机智能化分级系统测定了猪肉机器视觉级别GRADE、H、S、L值。结果表明：感官评分越高，H、S值越大，而L值越小。差异显著性分析表明1~6级各个感官级别之间的机器视觉GRADE和L值均存在显著性差异（P<0.05）；感官评分和机器视觉GRADE与H、S均呈极显著正相关（P<0.01），而与L呈极显著负相关（P<0.01）；加工特性指标中与色调值H显著相关性最高的是凝胶保水性，与亮度值L显著相关性最高的是凝胶强度，但与色饱和度值S相关性均不显著；利用机器视觉技术的L值可以预测肉糜的凝胶强度。 第三章：基于仪器法研究了加工特性与猪肉L*、a*、b*颜色变量之间的相关性，以及三种分级方法的关系和分级准确率。在机器视觉分级之后，立即通过仪器法测定猪肉的L*、a*、b*值。结果表明：感官评分越高，L*、b*值越小，而a*值越大；加工特性指标中与亮度值L*、a*值显著相关性最高的均是凝胶强度，与b*值显著相关性最高的是凝胶保水性； L*与感官评分存在高度显著相关，1~6级各个感官级别之间的L*值均存在显著性差异，说明L*可作为猪肉颜色等级划分的指标；利用仪器法的L*值可以预测肉糜的凝胶强度；感官评分、机器视觉级别和仪器法级别三者之间具有高度的显著相关性。若按6个级别划分，机器视觉技术和仪器法的总体准确率分别为65.33%和61.33%。若按3个级别划分，分别为83.61%和78.67%。 第四章：利用仪器法和机器视觉技术研究斩拌过程中肉糜加工特性的预测方法。以猪后腿肉和背膘为主要原料，按照肥瘦比3/7添加比例进行斩拌，分别于第6、7、8、9、10、12、14、16和18min取样，测定肉糜颜色变量和加工特性。结果表明：斩拌时间对肉糜的颜色、贮能模量(G’)、蒸煮损失、脂肪损失、硬度、弹性、出品率、Purge 损失均有显著影响。仪器法的肉糜L*值及机器视觉技术的L值与蒸煮损失、脂肪损失、弹性等加工特性之间具有较高的显著相关性，说明仪器法和机器视觉技术的亮度值均能很好地预测肉糜凝胶的蒸煮损失、乳化稳定性和凝胶特性。斩拌8-10min即可获得良好的凝胶特性。

As one of the most visual for consumer and critical parameters in pork quality, pork color determined the consumer’s intention of purchase. Pork color has been effected by gene types, raising conditions, different seasons, pre-slaughter treatment and other factors. At present, pork color grading had not been widely carried out in slaughter industry. Pork color grading had not been used by slaughter industriest, and pork had not reasonably been made for products processing, which caused a lot of losses. The thesis was mainly about pork functionalities based on sensory evaluation, Instrumental Method and Computer Vision System, found their predicting indcxes about functionalities. It initially eastablished their predictions by Instrumental Method and Computer Vision, and studied on prediction of chopping endpoint of meat emulsion. This study offered guidance of technology and theory about pork for pork producers, slaughter houses, meat processing industries, consumers. Chapter 1: 180 longissimus dorsi(LD) samples from pig(30 duplicate for each pork color grade) have been taken from pork carcass of normal slaughtering procedure, and took sensory evaluation. Then pork total protein solubility, gel strength (Rupture force) and elasticity, gel WHC (Water Holding Capacity) were tested. The results showed that all the functionalities indexes were extremely significant effected by pork color and ionic strength(P<0.01).The high ionic strength groups’ functionalities were more than the low groups’ (P<0.05). Gel strength was significant effected by interaction effect between sensory grade and ionic strength (P<0.05). However, total protein solubility, gel elasticity and gel WHC suffered no significant effect from interaction effect between sensory grade and ionic strength (P>0.05). sensory grade, ionic strength had respectively significant positive correlation total protein solubility, gel strength, gel elasticity and gel WHC (P<0.01). Chaper 2: After taking sensory evaluation, Computer Vison System measurement was once implemented. The results showed that the higher sensory grades was the larger H, S value, and the smaller L value. GRADE and L value was significantly different among the sensory grading groups of 1~6(P<0.05). sensory grades and GRADE had extremely significant positive correlation H, S value, and had extremely significant negative correlation L value. Gel WHC of functionalities indexes had the highest correlation H value, gel strength had the highest correlation L value. However, S value had no significant correlated on these functionalities. Gel strength of pork homogenate was predicted used as L value by Computer Vision System. Chapter 3: After measuring by Computer Vison System, colorimeter measurement by Instrumental Method was once implemented. The results showed that the higher color grades was the larger a* value, and the smaller L*, b* value. Gel strength of functionalities indexes had respectively the highest correlation L* value and a* value, gel WHC had the highest correlation b* value. L* value of color variables had the highest correlation sensory grades, and L* value was significantly different among the sensory grading groups of 1~6(P<0.05), which indicated that L* could be used the parameter of classification of pork color. Gel strength of pork homogenate was predicted used as L* value by Instrumental Method. Any two’s correlation coefficient from sensory grades, GRADE and color grades of Instrumental Method had significantly the higher correlation. If divided as 6 grades, Computer Vision System and Instrumental Method’s total grading ratio were respetively 65.33% and 61.33%, If divided as 3 grades, they were respetively 83.61% and 78.67%. Chpter 4: Prediction of the chopping endpoint of meat emulsion during the chopping process was studied by Instrumental Method and Computer Vision. Pork hindquarter and backfat were used to chop for meat emulsion with the fat/lean ratio of 3/7,and all the ingredients were chopped for 18min, During chopping, emulsion samples were obtained at each control time (at time intervals of 1min from 6min to 10min and then at time intervals of 2min up to 18min). The color and processing properties of meat emulsion were determined. The results showed that chopping time influenced color, Storage Modulus (G’), cooking loss, fat loss, hardness, springiness, yield and Purge loss. Moreover, during the chopping process L* and L value of meat emulsion was significantly correlated with the processing properties indexes, such as cooking loss, fat loss, springiness, and their correlation coefficient were relatively large, which meaned the lightness value of Instrumental Method and Computer Vision used predicting cooking loss, emusion stability and gel properties of meat emulsion. Under this experiment, when chopping time was 8-10min, gel properties of meat emulsion was better.