Human Body Data Measurement And Optimization Based On Kinect

The body height,chest circumference,waist size,and three-dimensional human body model are important data for the garment production industry.At present,the garment production industry obtains these data mainly relying on the traditional manual measurement method.This method is complicated to operate and has low efficiency,and is not suitable for large-scale human body size data collection.Tooling-based clothing companies need to collect a large amount of human body size information,and non-contact automatic measurement of human body data can quickly and accurately obtain human body data,which is important for improving the efficiency and measurement accuracy of large-scale human body sizes in the field of tooling production.The significance.Kinect equipment can obtain human body surface depth data and human skeleton data through structured light.Based on these information and computer image processing related theory,human body height,chest circumference,waist circumference and other human body size data can be measured and calculated.However,due to differences in body size,gender,etc.between different people,the mathematical model of the chest and waist is difficult to establish,resulting in a direct measurement of the body's bust through the Kinect,and the waist circumference has a large error.In this paper,we use human skeletal points to locate various parts of the human body and map the skeletal points to the depth image.The actual distance information represented by the skeletal points and the pixel information of the depth image are used to obtain the true distance represented by each pixel on the image plane.The depth information of the depth map of the human body is used to obtain the true distance of each pixel in the three-dimensional space.Then the number of pixels in the corresponding part of the human body is accumulated,so as to obtain the size information of the part.At the same time,due to factors such as equipment and environment,there will be certain errors in the measured human body size data.In human anthropometry,there is a certain proportional relationship among various parts of the human body.The ratio of the human skeletal segment and the measured height is obtained through the pixels of the depth image and the bone height is measured.The data acquires the real distance between the skeletal segments,combines the height data calculated from the constraints between the dimensions,and uses the least-squares method and iterative method to optimize the height measurement results to further improve the accuracy of the measurement data.This project implements the measurement system of human body data based on Kinect in the windows environment,and carries out data acquisition for multiple human bodies with different characteristics.The human body size calculation method proposed in this paper has good applicability to different groups of people and avoids the errors caused by different body shapes.At the same time,the measured size data is optimized to improve the accuracy of the human body size data.Because this method is an automatic measurement method,it can greatly improve the measurement efficiency and accuracy,and has a good practical value.