Search

The body temperature that can be used to determine the health of an animal is an important physiological feature of the animal. As the scale of modern poultry farming continues to expand, the use of traditional thermometer temperature measurement methods to determine its health status is no longer applicable. The rapid development of thermal infrared imaging makes it possible to quickly measure the temperature, but whether the measured surface temperature can characterize the health of the poultry is uncertain. In order to determine the health status of laying hens through body temperature, the surface temperature and true body temperature of healthy laying hens were measured at different ambient temperatures. A method for extracting surface temperature of non - feather area based on temperature threshold is proposed. On this basis, the prediction model of body surface temperature was obtained by regression analysis. By comparing the measured temperature of the pathological individual with the predicted temperature, the discriminant threshold of the pathological individual was obtained. Fifty healthy chickens with healthy and pathologic pathology were selected to determine their health status by using the above method. The correct rate was 85.66%. The results show that the health status of poultry can be judged by applying the thermal imager to measure the surface temperature.

Shape is one of the most important indexes for fruit quality evaluation, and extracting the shape characters of fruit is the key step for the shape grading. However, it is very difficult to get the accurate description of fruit shape because of a variety of fruit shape in different attitudes. In this research, 4 images were taken for each apple respectively. There are 144 images for 36 apples, in which 12 apples being considered to have the best shape, and 11 ones having big deformation, and the other 13 ones having small deformation. Then process these images and get 245 shape character parameters of each boundary of the fruit through signature description, Hu moments, Zernike moments, wavelet moments, Fourier description, and wavelet description. This paper is trying to get efficient and practical shape indexes and find a feasible way for describing fruit shape.