Infrared image enhancement method for vehicle-mounted infrared pedestrian detection

Abstract

The invention discloses an infrared image enhancement method for vehicle-mounted infrared pedestrian detection. The method comprises the steps that 1, an image processor carries out the adaptive median filtering of an infrared image, and removes the particle noise and salt and pepper noise in the infrared image; 2, the image processor performs equalization processing on the infrared image processed in the step 1 by adopting a platform histogram equalization method to obtain an equalized image; 3, the image processor performs gamma transformation on the image equalized in the step 2 to obtain an image y1; 4, the image processor firstly performs Laplace sharpening processing on the infrared image processed in the step 1, and then performs non-maximum suppression processing to obtain an imagey2; and 5, the image processor performs weighted combination processing on the image y1 obtained in the step 3 and the image y2 obtained in the step 4 to obtain an image y. The method is novel and reasonable in design, can improve the enhancement effect of the infrared image, obtains the overall information of the image, is high in recognition degree, is rich in layering, is high in practicality,and is high in popularization and application values.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an infrared image enhancement method for vehicle-mounted infrared pedestrian detection. Background technique [0002] In order to protect pedestrians in the traffic system, reduce the driver's operating intensity, learn the movement status of pedestrians, and provide safety warnings for drivers in advance, so that drivers can respond in time to avoid, vehicle-mounted pedestrian detection has received widespread attention. Traffic accidents are of great significance. According to the different working environments of the vehicle-mounted camera that collects images, it can be divided into: pedestrian detection based on visible light images and pedestrian detection based on infrared images. The acquisition of visible light images is easily affected by light, and the accuracy of pedestrian recognition is reduced, and ordinary cameras cannot be used at night, in ...

AI Technical Summary

Problems solved by technology

However, due to the limitation of infrared thermal imaging technology and the influence of radiation factors such as noise and atmosphere, the vehicle-mounted infrared image presents the problems of concentrated gray scale distribution, low contrast, and poor resolution. Therefore, how to improve the enhancement of vehicle-mounted infrared image effect becomes an important research topic

[0003] In recent years, many scholars at home and abroad have proposed many infrared image enhancement methods based on this. For example, Liu Yitong published "Based on Kinect and Histogram Equalization" on pages 173-176 of "Ship Science and Technology" in Volume 37, Issue 3, 2015. In the article "Infrared Image Enhancement Algorithm", the Kinect and histogram equalized infrared image enhancement method are applied to the infrared image of ships in complex sea areas. The method first obtains the depth image of the infrared image through Kinect, and according to the different depth of field, the ship and The background area is separated; then the initial threshold is set, and after redundant grayscale compression, power index stretching and histogram equalization, the pixels of the image are arranged at equal intervals on the corresponding grayscale, and the image contrast is improved, but in In the pedestrian environment when the background is bright, the enhancement effect is not soft, rather stiff, unable to distinguish the pedestrian from the background; In the article "A New Adaptive Infrared Image Enhancement Algorithm" published on pages 768, 769, and 776, a new adaptive infrared image enhancement algorithm is proposed, which performs adaptive double thresholding on the histogram of the original image. Mapping and under-sampling projection processing realize the redistribution of image gray levels; on the basis of reducing redundant gray levels, the dynamic range occupied by the target part is effectively widened, and the contrast between the target and the background is improved. However, in the infrared In the pedestrian image, the target is small, the background is large, and the image enhancement effect is not ideal

Images