Research on an automatic counting system of soft fibrils collection

An automatic counting system on soft fibrils collection was designed. The collection slice was manufactured by a self-designed slicer, imaged by a microscope optical system, and captured by CCD (Charge coupled device). After that the slice image was transferred into a computer by the image acquisition card. In the system the hemispherical LED source which had separate switches was used. The intensity of the LED source could be adjusted independently. The quantities of the collection could be got by the image processing algorithm proposed in the paper. The binary image could be obtained based on maximum region entropy, SOFM (Self organization feature map) neural network and image optimal threshold segmentation. Based on the statistics the boundary of the binary image was separated further. The number of the collection was calculated through the counting algorithm. The experimental results show that the system detection error is no more than 1%; the system has the much higher repeatability and precision.