Image Identification And Feature Point Calculation Based On HOUGH Transformation To The Known Figure

The machine vision is used to simulate human's visual function with the computer via useful information in the objective image mainly, carries on processing and performing to understand and compute. Finally uses in examining, surveying and controlling actually, then the system uses the machine vision in realizing automatic device control, completing the industry packing work task. The system includes the photo source (is natural photo source), the optical system, the image capture system, the image digitization module, the digital image processing module, the intelligent judgment decision-making module and the machinery carrying out module and the history data record, and the inquiry module. Picture signal processing and the actual addressing algorithm is the machine vision system core, it's equal to human's cerebrum. How process and operate the image, namely the algorithm fully manifests here, this is the machine vision system development's key point and the nodus. This project has also carried on the research and development as the key point, the image process and the coordination point algorithm have also be carried on the detailed expounding in the article. The aim of the article's studies lies in enhances the imagery processing algorithm speed, the precision as well as the recognition rate. Enhancement of the imagery processing speed may strengthen the control system greatly at times, enhances the production efficiency. Precision and recognition rate enhancement improvement control accuracy, fuller display mechanical vision superiority, completes high accuracy computation which the human eye all is even unable to realize, thus improvement working conditions and work completion quality. This article researched traditional HOUGH transformation, uses the gradient edge examination method in its foundation, once more reduces the HOUGH transformation computation quantity through the distinction image function edge change direction the operand, enhance the mechanical recognition speed. The system software realization is the key content in this article, the system must act according to the customer demand, depending upon the current hardware disposition and the lower position machine communication cooperation to realize following function: the automatic control moving the material vessel to the assign position; automatic addressing, control charging device to target point; examines material data at times, automatic stopping filling material; automatically sends out material vessel, circulation work; tracks record the material data and each kind of condition; Historical data detailed inquiry and statistics; man-machine interactive contact surface which be friend and easy to operate, Is advantageous for the user operation, starts as necessary and stops, processes each kind of situation. This software system in order to can realize real-time image data gathering, real-time intercepts work and so on, adopted the multithreading, the multi-duties parallel movement. Mainly includes the main-control system, the image gathering system, the target point recognition system, the database system, the pronunciation alarm system: 1,The main-control system main-control system's main realization: coordinated the relation and correspondence among image thread, input output thread, serial thread , and the overall variable, namely thread synchronization control; input gain; outputcontrol. Input gain: The main-control system detect the 16 digital signal input at times, according to various needles foot switch quantity condition, distinguishes the signal the input situation. The main-control system examines the condition that each needle foot directly using the C++ position section operation, it is advantageous for the operation. All input signal uses the low level input, is advantageous for the monitor input device working mode active status, Namely the input device appears when the breakdown (constantly: High level; has the signal: low level. When should not have the input signal examines to the low electricity proposition malfunction), the main-control system may examine as necessary and takes the corresponding processing measure. This system worked in bad environment, high efficiency electrical machinery magnetic field disturbance commonly. In the gain input signal process adopts bidirectional disappears shakes filter, avoids the input signal gain mistake. Output control: the output board card and the input board card is actual the same one. Also uses 16 digital signals output, uses the high level output completely. The digital signal outputs 24V signal to the enlargement module enlarges after the light partner isolation to give the solenoid valve, control output control system and all other charging device movement. 2,Image gathering system This system has considered the image gathering speed emphatically during the design process. Because this project is the multi-locations joint operation, but the image gathering card uses at the same place, must carry on the gathering card source road perform handover in the system movement process, the image gathering card when carries on the source road cut as a result of image gatheringcard hardware limit, must waste some time, therefore, adopts image gathering and the processing multi-tasking way, namely: when the gathering card cuts to the location 1 source road, processes the location 2 image data; when the gathering card cuts to the location 2 sources roads, processes the location 1 image data. Like this uses in the source road cut machine-hour the imagery processing, enhanced the system running rate and the efficiency in the certain degree. 3,The target point recognition system In this system, work center and the key point are the HOUGH transformation improvement, enhances the running rate and the efficiency. At the same time, in order to increase the target point computation precision, after the image HOUGH transformation, this system increased the nature heart operation step. Through histogram way true standard quality heart algorithm gradation threshold value, further more increased the target point physical location precision. The traditional HOUGH transformation must satisfy the specific condition in the HOUGH space around the spot, carries on the known graph perimeter points to make the mark, but uses the directional derivative, only in satisfies a specific condition certain direction and certain distance place coordinates from the HOUGH space counts, this may reduce its operand greatly, thus enhancement operating speed. 4,Database system In this system, we considered that enhance the operation as far as possible. In the database modeling process the convenience and reduces the floppy disk space to take, the inquiry speed as far as possible is as far as possible quick. The chronometer saves the different type data, divides feeds in raw material the datadetailed table, the general data sheet and the wrong diary table and so on. In the insertion, the deletion, the renewal, the inquiry data watch with the table carries on the link through the outside key. 5th, pronunciation alarm system of this project is more direct-viewing human nature for the realization reports to the police the control, uses the sounds of people simulation pronunciation system to carry on reports to the police the prompt function. All reports functions are completed by the computer sound production, reports to the police the prompt promptly, more accurate, is more direct-viewing. The newest pronunciation storehouse and the pronunciation module which provides using Microsoft Corporation, strengthened the procedure probability, reduced the development cycle; this system may carry on English, Japanese, Chinese and so on the many kinds of languages reports to the police, strengthened the software adapting environment and the flexibility.