<![CDATA[In the road traffic space which is often encountered by passing traffic type of vehicle. To find out the traffic conditions that are needed to calculate vehicle traffic, such as using counting or recording CCTV video. This continues the long and long process that was completed on the error data and the slow pace of traffic engineering decisions. This method is difficult to do in full because of the limited number of counters. This can be done by involving digital processing and CCTV video to be able to classify and transfer vehicle type objects. There are several methods for sharing object imagery, such as SIFT, edge detection and Monte Carlo. This research tries to use the Background Substraction and Blob Detection methods because of its superiority in determining objects and backgrounds and being able to maintain moving objects as well as analyzing screen area calculations. The results of testing with this method obtained the MSE value at the threshold of 100 and 3x3 kernel filter with a pixel area of motorcycle 34-63 pixel-X, 67-155 pixel-Y and cars 73-200 pixel-X, 79-307 pixel-Y and bus / truck 130-128 pixel-X, 305-376 pixel-Y. On evaluation, use the confusion matrix obtained in the morning with an average total of 92% and at night with a total average of 73%. It can be concluded by using CCTV installation parameters and the method used can yields higher accuracy in the morning than at night with the weakness of compiling objects that can make it easier to make objects and test the night to obtain light from vehicle lights generated as vehicle objects the flight.]]>
