Zebra crossing detection algorithm applied for intelligent drive

Abstract

The present invention relates to a zebra crossing detection algorithm applied for intelligent drive, comprising: capturing an original gray image and extracting a zone of interest; preprocessing the extracted zone of interest so as to obtain a binary image; traversing from the first line of the binary image, extracting gray values, storing them into an array and transforming the values in the array into a semaphore; judging whether the semaphore corresponding to the present line is a square signal, if so, the semaphore is the semaphore extracted from the zebra crossing and the present line is the starting line of the zebra crossing, otherwise the semaphore is not the semaphore extracted from the zebra crossing; finishing traversing if the line is the last line of the image; and checking a flag bit of the zebra crossing, if the flag bit is 1, a zebra crossing is present in the image, and if the flag bit is 0, no zebra crossing is present in the image. By using the texture information and black and white color information of the zebra crossing, the algorithm in the present invention transforms two-dimensional gray judgement of the zebra crossing into one-dimensional semaphore judgement of the zebra crossing, reduces the dimension for data processing and increases the speed and accuracy for identifying a zebra crossing.

Description

technical field [0001] The invention relates to the field of intelligent driving, in particular to a zebra crossing detection algorithm applied to intelligent driving. Background technique [0002] With the development of society, cars play an increasingly important role in family life. However, the huge car base has brought huge hidden dangers to road safety issues. Especially at traffic intersections, accidents occur frequently especially. The detection of road markings at traffic intersections is imperative. As for zebra crossing, as an important part of traffic intersection signs, the accuracy and real-time performance of its detection are particularly important. [0003] At present, the detection of zebra crossing mainly utilizes the texture information and color information presented by the zebra crossing in the image. The use of texture information mainly means that the zebra crossing presents an approximately parallel characteristic in the image. The color infor...

AI Technical Summary

Problems solved by technology

[0005] Although the bipolar coefficient method can detect whether there is a zebra crossing on the target image, because it needs to calculate the bipolar coefficient of the specified "pixel block" in the image, it also needs to slide the position of the pixel block to calculate the value of the bipolar coefficient in different areas, resulting in The calculation is complex and the calculation speed is slow, which cannot meet the needs of the intelligent driving real-time system

At the same time, because the position of the zebra crossing cannot be given during the calculation process, the practicality is limited to a certain extent.

Images