The invention provides a Prewitt operator-based pavement crack recognition method and system. The method comprises the steps of S1, performing first image processing on an input pavement image to remove a region outside a pavement so as to obtain a pavement region image; S2, performing second image processing on the pavement region image to remove a pavement boundary line; and S3, performing preliminary recognition and accurate recognition on a pavement crack to obtain pavement crack data. According to the method and the system, the pavement image is subjected to calculation processing based on a Prewitt operator, a useless image outside the pavement is filtered, image interference and noises of a pavement region are gradually removed, the pavement crack is extracted, the length of the crack is calculated, and quantitative output is performed, thereby facilitating comparison with an actual standard; the method and the system are mainly applied to the pavement crack detection of complex high-altitude motion background and scene; and compared with various common crack detection algorithms, the method and the system have the advantages that the adaptability is higher and a better aerial image-based crack recognition effect is achieved.

Prewitt operator-based pavement crack recognition method and system

The invention discloses an unmanned aerial vehicle for emergency commanding and lane occupation evidence taking. The unmanned aerial vehicle is provided with forward binocular cameras, a backward binocular camera, a forward binocular image stitching module, a forward monocular vehicle identification module and a backward binocular image stitching module, wherein the forward binocular image stitching module and the forward monocular vehicle identification module correspond to the forward binocular cameras, and the backward binocular image stitching module corresponds to the backward binocular camera. The forward binocular cameras include the forward left ocular camera and the forward right ocular camera. The forward binocular image stitching module is used for stitching video images of the forward binocular cameras and outputting a forward stitching image and a forward right ocular image. The backward binocular image stitching module is used for stitching video images of the backward binocular camera and outputting a backward stitching image. The monocular vehicle identification module is used for identifying vehicle attributes according to the video image of the forward right ocular camera. The unmanned aerial vehicle can perform illegal detection and evidence taking on vehicles occupying an emergency lane, and therefore a traffic jam can be relieved effectively in time.

Unmanned aerial vehicle for emergency commanding and lane occupation evidence taking

The invention discloses an expressway lane line detection method based on an onboard camera. The expressway lane line detection method comprises the following steps: 1: reading an image by frame; 2: extracting an image of a region of interest; 3: carrying out downsampling on the image of the region of interest; 4: on the basis of a colorful image, carrying out binaryzation to obtain a binary image which contains a lane line; 5: utilizing lane line width information to remove chaff interferents including vehicles, roadside buildings and the like, then, carrying out filtering according to the area of each connected domain in the image, and removing desultory points with small areas to obtain a lane line image; 6: obtaining a line segment parameter; 7: according to lane line distribution characteristics, ejecting a non lane line straight line; and 8: updating the width and the coordinate of the left upper corner of the ROI (Region Of Interest). The expressway lane line detection method based on the onboard camera can detect all lane lines in a video image, and provides an important guarantee for the illegal behavior of each vehicle on the expressway. The method is high in accuracy and robustness.

Expressway lane line detection method based on onboard camera

The invention provides a method for detecting an illegal vehicle of a high-speed emergency lane by using an unmanned aerial machine. The method comprises the following steps of updating an acquired initial image of a detection area by using a flood fill method; then conducting detection of a salient area of the illegal vehicle on the updated image of the detection area based on the visual significance analysis, and obtaining a saliency map of the detection area; judging whether or not the illegal vehicle exists in the detection area based on statistical results of pixel points in the saliency map; subsequently according to the standard deviation change of a detection frame, calculating the illegal parking time of the illegal vehicle, and finally storing vehicle violation information obtained in the above two steps into a high-speed traffic management system. The method for detecting an illegal vehicle of a high-speed emergency lane by using an unmanned aerial machine can utilize the unmanned aerial machine to conduct automatic detection and snapshots for evidence investigation on vehicles which illegally occupy the emergency lane, and can carry out efficient and real-time monitoring of the emergency lane, thereby greatly enhancing the control intensity of the traffic police on illegal behaviors on a highway.

Method for detecting illegal vehicle of high-speed emergency lane by using unmanned aerial machine

The invention discloses an image reconstruction model training method and an image super-resolution reconstruction method and device, and belongs to the technical field of image super-resolution. Themethod comprises the steps of obtaining a sample set by preprocessing an image, establishing an image reconstruction model for image super-resolution reconstruction; using the sample set to train andtest the image reconstruction model; in the image reconstruction model, using a feature extraction network for performing feature extraction on a low-resolution image and inputting into a first residual network, wherein the m cascaded residual networks are respectively used for carrying out feature extraction on an output image of a previous network and then superposing the output image with the image, m attention networks are respectively used for extracting images of a region of interest from the output images of the m residual network, and an amplification network is used for fusing and amplifying the output images of the attention networks and the m residual networks, so that the output images and the image subjected to bicubic interpolation amplification are fused by the first fusionlayer. According to the present invention, the visual effect of the reconstructed image can be effectively improved.

Image reconstruction model training method and image super-resolution reconstruction method and device

The invention discloses a single image super-resolution reconstruction method based on a depth component learning network, which comprises the following steps: expanding a training sample image and carrying out region extraction and degradation operations to obtain corresponding high-resolution and low-resolution image training sets; a deep network with component learning structure being constructed. The network decomposes the input low-resolution image into global components, and then predicts the corresponding image in high-resolution space using the residual components extracted from the global components. Batch random gradient descent method and back propagation algorithm are used to train the constructed deep component network iteratively on the training set, and the model with optimized weights is obtained. Reconstruction of Low Resolution Images Using the Trained Component Networks; the reconstruction result is restored to the original color space, and the final output of the super-resolution reconstruction is obtained. The method of the invention not only can improve the quality of the reconstructed super-resolution image, but also can improve the operation speed of the model.

Single image super-resolution reconstruction method based on depth component learning network

The invention relates to an expressway vehicle detection method based on improved visual background extraction, which comprises the steps of 1, initializing a read-in image, wherein an input color image is converted into a grayscale image; 2, carrying out background molding on the image, carrying out background initialization if the image is a first frame image, and otherwise, carrying out background updating so as segment out the background; 3, segmenting out a road surface from the foreground; 4, carrying out eight-neighborhood filling on the image; 5, removing a lane line by using a characteristic that the lane line is narrow; 6, filling the internal part of a vehicle by using a longitudinal filing method; and 7, extracting a vehicle area and drawing an external rectangular frame of the vehicle area. According to the invention, vehicles on the expressway in aerial video of an unmanned aerial vehicle are detected by adopting a mode of image processing, the cost is low, the detection precision is high, the real-time performance is good, and the application range is wide.

Expressway vehicle detection method based on visual background extraction

The invention discloses a single image super-resolution reconstruction method based on two-way alignment sparse representation, comprising the following steps: (1) reading a color low-resolution image, and transforming the image from an RGB color space to an YCbCr color space; (2) primarily converting the image into an image of a target size through bicubic interpolation, performing a super-resolution reconstruction modeling operation based on two-way alignment sparse representation on the brightness component of the image after conversion, and iteratively solving the reconstruction model through an iterative shrinkage thresholding algorithm to get the optimal estimated value of the brightness component of the high-resolution image; and (3) transforming the image from the YCbCr color space to the RGB color space, and getting the final output of super-resolution reconstruction The method not only can improve the quality of reconstructed images, but also is superior to the traditional methods in robustness

Zeng Weili

Papers

Single image super-resolution reconstruction method based on depth component learning network

Expressway vehicle detection method based on visual background extraction

Expressway lane line detection method based on onboard camera

Single image super-resolution reconstruction method based on two-way alignment sparse representation