Research On Calorimetric Measurement Technology Based On Fruit And Vegetable Identification

With the continuous development of artificial intelligence field,people’s life is gradually stepping to intellectualization,among which image recognition technology occupies an important position in artificial intelligence.With the improvement of people’s living standard and the enrichment of food categories,how to understand personal dietary pattern,nutrition and health condition,and improve nutritional intake and obtain healthy lifestyle becomes a matter that people pay more and more attention to.At present,there is nutrition and calorimetric measuring system aiming at western-style food in foreign countries,however,because of the complexity of cooking process and the variety of food categories in Chinese meals,there is no nutrition and calorimetric measuring system aiming at Chinese food yet.In this paper,the research is carried out aiming at how to measure the nutrition intake of Chinese meal,because of the complexity of the cooking process of Chinese meal,the research in this paper is focused on the identification and nutrition measurement of fruits and vegetables part in the raw materials of Chinese meal and the overall calories and nutrition estimation system.First,in this paper,the fruits and vegetables positioning technology of fruits and vegetables images in complicated environment and uncontrollable conditions is studied,applying improved manifold significance test ranking,fruits and vegetables positioning is implemented with the integration of gradient energy and positional information based on original technology;then,the color features,contour features and textural features of fruits and vegetables are integrated in this paper,SVM classifier is improved through introducing TPS transformation model,and the integration features are classified with the improved SVM classifier;in this paper,data set is tested with convolutional neural network and improved SVM classifier,the experiment result shows that the improved SVMclassifier has higher accuracy for the classification result of small sample data;finally,thumb measurement is applied in this paper,through introducing adult thumb as the measuring scale,the fruits and vegetables volume is estimated and volume data is obtained,and meanwhile,fruits and vegetables nutrition and calories information is obtained through the database of China Food Nutrition Network.Research on calorimetric measurement technology based on fruit and vegetable identification can realize real-time fruits and vegetables category testing and automatically display the calories and nutrition content of the current fruits and vegetables,and provide reasonable nutrition improvement suggestions.The system can make the users voluntarily conclude the daily nutrition and calories intake,which is good for improving the dietary pattern and health condition of people and is of broad application prospect as well.