In the course of wheat breeding and cultivation trials, the counting of wheat grains is a very important step. In the past, manual counting was time-consuming and labor-intensive, and the workload was very large. In order to improve efficiency, people invented an automatic seed counting system. The automatic seed counting system first uses electromagnetic vibration to arrange the grains into a single grain queue, which is fed into the photoelectric conversion tank to form a photoelectric pulsation, and then the number of grains is obtained by the counting circuit.
The key to using the seed automatic counting system is the image segmentation algorithm. At present, the more common types of algorithms are the following types: (1) Corrosion-expansion method. The target image is gradually eroded by a set structural element, and several layers are gradually stripped off. . During the process of corrosion, the adhered grains are separated into a single object, and then the partitioning boundary is found after bulging. (2) Distance conversion and watershed algorithm. Distance images are obtained by performing distance conversion on the binary image. The grain boundary is obtained by using the watershed algorithm to complete image segmentation. (3) Corrosion expansion and active contour model are first used. A seed area is obtained inside the grain, and its boundary is taken as the initial deformation curve of the active contour model. Then, an active contouring model is used to reconstruct the grain outline to obtain an accurate grain number.
Due to the large sample size of wheat grains, the high requirement of counting speed and accuracy, the accuracy of the watershed algorithm is not sufficient, and the calculation of the active wheel phoenix model is costly and slow, and it is difficult to apply. In contrast, the traditional extreme corrosion method is still the most effective method, but it needs to be improved according to the specificity of wheat grain. The wheat grain shape is close to an ellipse, and there is a statistical consistency between the size and shape of the grain between the same sample, which are the basis for the improved algorithm.
The key to using the seed automatic counting system is the image segmentation algorithm. At present, the more common types of algorithms are the following types: (1) Corrosion-expansion method. The target image is gradually eroded by a set structural element, and several layers are gradually stripped off. . During the process of corrosion, the adhered grains are separated into a single object, and then the partitioning boundary is found after bulging. (2) Distance conversion and watershed algorithm. Distance images are obtained by performing distance conversion on the binary image. The grain boundary is obtained by using the watershed algorithm to complete image segmentation. (3) Corrosion expansion and active contour model are first used. A seed area is obtained inside the grain, and its boundary is taken as the initial deformation curve of the active contour model. Then, an active contouring model is used to reconstruct the grain outline to obtain an accurate grain number.
Due to the large sample size of wheat grains, the high requirement of counting speed and accuracy, the accuracy of the watershed algorithm is not sufficient, and the calculation of the active wheel phoenix model is costly and slow, and it is difficult to apply. In contrast, the traditional extreme corrosion method is still the most effective method, but it needs to be improved according to the specificity of wheat grain. The wheat grain shape is close to an ellipse, and there is a statistical consistency between the size and shape of the grain between the same sample, which are the basis for the improved algorithm.
Gate Valve,Slide Gate Valve,4 Inch Gate Valve,Stainless Steel Gate Valve
Wuzhong Instrument Co., Ltd. , http://www.wzivalve.com