Car dashboard computer vision detection system
|School||Changchun University of|
|Course||Mechanical Design and Theory|
|Keywords||Car dashboard Computer Vision Image processing Signal source driving Hough transform|
The car dashboard concentrated the entire vehicle instrument display the working condition of the car . Different car dashboard instrument vary , but conventional instruments of the general car has a speedometer , engine tachometer , water temperature gauge , fuel gauge , a variety of signal alarm lights and lighting , etc., before entering the vehicle , need to conduct a final inspection quality . -Like detection methods are manual detection meter pointer indicates error , to determine whether the alarm lights and lights for damage , missing equipment or other appearance defects , this approach by subjective factors such as operator observation angle , distance and the degree of fatigue , error and poor reliability. Instead of artificial computer vision to the dashboard automatically detect various errors caused by manual inspection can be overcome to improve detection accuracy and efficiency . Relying on technological development projects in Jilin Province (No. 20080535 ) in auto dashboard computer vision detection research papers , design of automatic detection system based on computer vision detection technology car dashboard . Thesis design work includes : the design of the source drive vehicles each instrument so that the meter pointer rotation , shooting dynamic car dashboard image stored in the computer for image processing ; using median filtering method to remove noise in the image pre-processing , using Laplace operator sub - edge detection , the use of the otsu methods meter image binarization ; using image morphological transformation, proposed closing operation after first use refinement algorithm , experiments show that this method can effectively remove the interference image acquisition ; using Hough transform algorithm for process optimization , extraction of the straight line of the automobile instrument pointer , to improve the detection accuracy and processing time of the dashboard instrument pointer ; through the angle method, and the comparison of the distance method , determined using the readings of the distance method automotive instrument pointer , and further improve dashboard instrument pointer detection accuracy. Experimental results show that , In this paper, the design of automobile instrument pointer extraction algorithm has strong anti-interference ability and stability . Online testing using computer vision technology for automotive instrument will gradually replace manual inspection , automotive instrument detection industry mainstream .