Development Of Capture System Targeting Cell Under Microscope

Computer applications are nothing than common in various fields owe to the rapid development of computer technology,among which computer vision and image process ing play a pivotal key especially in biomedical science.Researchers had no choice but to conduct experiment on cells by observing cells in early time.Though semi-automated method alleviated manual works to observe glass slides at certain level,there are however,stills lots of aspects to be improved.This article proposes a new method to cap ture cells and perform initial automatic detection and observing sheet by computer image processing.In the first part,this article illustrated how the microscopic vision images are converted into computer and to digital images.Some adjustments such as gradation processing,smoothing processing and so on conducted to fix problems existing in the images.Several prevalent methods of edge detection introduced in second part.Edge detection is the basic step yet the key step in the image recognition and processing.A cell is determined to be a real cell or not by its edge.We choose Canny edge detection algorithm for a better performance after comparison with other widely used algorithms.Hough transformation is used frequently in common pattern recognition.However,it's time consuming when it is applied to identify the cells in an image which is complicate.So we choose RED algorithm,which slashes computing time and improves detection efficiency greatly,instead.Finally,we get its coordinates by identification of the cell,and send computations result to lower computer.After analyzing the results,lower computer sends command to stepper motor,which connects the shaft of bench,to adjust the targeted cell to the center of scope.Our method proposed in this article achieves cell seeking and recognition,and meets the requirements basically.