18results about How to "Rich image information" patented technology

The invention provides a single target tracking method based on aggregation token perception features, and particularly relates to the technical field of vision and image processing. A single-target tracking method based on aggregation token sensing features comprises the following steps: step 1, under single-target tracking based on the aggregation token sensing features, constructing an overall model composed of a ViT-B backbone network, an aggregation token sensing feature module, a multi-scale feature fusion module and a positioning head; 2, initialization: giving a data set X including N video sequences, selecting a first frame of image and an initial bounding box position of a target, using a target detection model of a pre-trained ViT-B backbone network or a manually labeled bounding box to initialize a tracking model, feature extraction and enhancement process, obtaining template key information through an aggregation token perception feature module, and obtaining template key information; and utilizing a deep neural network to accurately extract the feature representation of the current frame image and the target area, and inputting the cut target area into a feature extraction network.

An optical lens, a camera module and a terminal device are provided, the optical lens has six lenses with refractive power, the six lenses comprise a first lens to a sixth lens in sequence from an object side to an image side along an optical axis, the refractive power of the first lens is negative, negative, positive, negative, positive and negative, and the object side surface and the image side surface of the first lens are a convex surface and a concave surface near the optical axis respectively; the object-side surface and the image-side surface of the second lens element are respectively a concave surface and a convex surface near the optical axis; the object side surface and the image side surface of the third lens are convex surfaces near the optical axis; the fourth lens element has a convex object-side surface and a concave image-side surface in paraxial regions, respectively; the fifth lens element has an object-side surface and an image-side surface being convex in a paraxial region. The sixth lens element has a concave image-side surface in a paraxial region. The seventh lens element has a convex object-side surface and a concave image-side surface in paraxial regions, respectively. The optical lens satisfies the following relational expressions: 65 degree lt; a FOVlt; the angle is 75 degrees, and the tilt is 3.5 degrees; tTL / Hlt; 4.2. The optical lens provided by the invention can realize miniaturization and has the characteristic of a large field angle.

The invention relates to the technical field of artificial intelligence, and discloses a self-adaptive indoor positioning and path optimization method based on artificial intelligence, which comprises the following steps: S1, collecting visual positioning data, deploying the position and angle of a camera to cover an indoor area, then collecting image data of an indoor environment in real time, and setting the position and angle of the camera to be the same as the position and angle of the camera; calculating and determining the positions of the target object and other obstacles in the image through visual artificial intelligence; s2, collecting wireless Bluetooth positioning data; s3, positioning data fusion and evaluation; s4, target setting and path planning; s5, executing and monitoring a path; and S6, dynamic path optimization. Visual artificial intelligence and wireless Bluetooth positioning are combined to improve positioning accuracy and stability, visual positioning provides abundant image information, wireless Bluetooth positioning ensures coverage and stability, and matching of the visual positioning and the wireless Bluetooth positioning can correct errors, adapt to different environment changes and provide an accurate position basis for path optimization. The robot navigation efficiency and the task execution success rate are improved.

The embodiment of the invention relates to a train number state video recognition method and device, equipment, a medium and a program product, and relates to the technical field of video information processing. The method comprises the steps of performing frame-by-frame reasoning on each frame of sequence image in a target monitoring video through a pre-trained train number state recognition model to obtain the state of a vehicle in each frame of image of the target monitoring video, and if the number of times of occurrence of the same vehicle prediction state before a current video frame in the target monitoring video reaches a preset threshold value, determining that the current video frame of the target monitoring video is the same as the current video frame of the target monitoring video. And determining the vehicle prediction state as the state of the vehicle in the current frame image, and executing corresponding video editing once according to the current frame image to obtain at least one complete vehicle number state video clip corresponding to the target monitoring video. Wherein the train number state recognition model is obtained by learning video time sequence feature training based on a 2D convolutional neural network structure. By adopting the method, the complete train number video can be quickly and conveniently identified.

The invention provides a Transform-based wound image enhancement method, which comprises the following steps of: simulating and adjusting related parameters in a camera internal ISP (Internet Service Provider) process in a mode of fully supervised training and dynamic inquiry learning, introducing a local mapping prediction image pixel-level multiplication graph and addition graph with an attention mechanism, mapping an image from a standard RGB (Red, Green, Blue) space to a camera RGB space, and carrying out image enhancement. According to the invention, image enhancement is carried out through an end-to-end Transform network, the problems of image degradation and the like caused by low illumination are solved, image enhancement is carried out on the problem of non-uniform illumination of a wound image caused by a shooting process, and the problems of image distortion and the like caused by non-uniform illumination are solved.

The utility model discloses a pressing device with a visual inspection function, which belongs to the field of pressing devices and comprises a bottom plate, a pressing worktable and a support are fixedly connected onto the bottom plate, a cylinder is fixedly connected onto the support, a pressing plate is fixedly connected onto an output shaft of the cylinder, and a visual inspection component is movably mounted on the bottom plate and the support. Through the visual detection assembly, after carpet workpieces are pressed, a first industrial camera shoots the front faces of the carpet workpieces, an adsorption box adsorbs the carpet workpieces and then turns the carpet workpieces to a reverse face collection table, and a second industrial camera shoots the reverse faces of the carpet workpieces, so that image information of the front faces and the reverse faces of the carpet workpieces is complete; the judgment result of the pressing effect is more accurate, the carpet workpiece is overturned to the reverse collecting table to be discharged, operation below a pressing plate is avoided, the discharging safety is higher, the first moving table and the second moving table are convenient to move, and the positions of the first industrial camera and the second industrial camera are more convenient to adjust.

The invention discloses an image processing method and apparatus, a device and a storage medium. The method comprises the steps of obtaining a to-be-matted target image and description information of a target background; the target background is different from the background in the target image; fusing the image features of the target image and the description features of the description information to obtain fused features; and decoding based on the fusion feature to obtain a target decoding result, the target decoding result comprising a foreground segmentation result of the target image and a fusion image of the foreground and the target background in the target image. In this way, rich image information can be provided for a user.

The present invention discloses an Autoformer sintered ore FeO content soft measurement method based on multi-source information fusion. First, in view of the complex characteristics of sintering process data such as dynamics, strong nonlinearity and multi-source heterogeneity, a multi-source information fusion strategy is proposed to realize feature-level information fusion and data-level information fusion, which are used for image feature extraction of sintered ore cross-section and multi-source heterogeneous data integration, respectively. Subsequently, more comprehensive sintering process information is input into the downstream Autoformer model in a serial manner. The above-mentioned serial connection not only improves the information capacity, but also provides prior information about the FeO content level of the sintered ore, which is conducive to the Autoformer capturing the complex time series distribution in the sintering process, so that the Autoformer model based on multi-source information fusion has more accurate modeling capabilities in real industrial scenarios.

The utility model discloses a multi-degree-of-freedom industrial X-ray tomography device, and relates to the technical field of imaging. The device comprises a base, a plurality of supporting legs are installed at the bottom of the base, and a supporting mechanism, an adjusting mechanism and a rotating mechanism are arranged on the base. The supporting mechanism comprises two reinforcing plates installed on the sides, close to each other, of the multiple supporting legs, supporting plates are installed on the left side and the right side of the top of the base correspondingly, and a first motor is installed on the left side of the corresponding supporting plate. According to the multi-degree-of-freedom industrial X-ray tomography imaging device, by arranging the supporting mechanism, the adjusting mechanism and the rotating mechanism, multi-angle scanning can be conducted on an object, and the imaging effect of the device can be improved.

The present invention relates to the field of municipal pipeline detection technology, and in particular to a method for real-time autonomous identification of typical defects in municipal pipelines. The method comprises the following steps: S1, collecting and acquiring original sonar data of the municipal pipeline environment through a mechanical scanning sonar mounted on a crawling robot; S2, real-time analysis of the original sonar data obtained in step S1 to obtain a data matrix containing beam angles and sound intensity values ​​of sampling points; S3, real-time generation of a mechanical scanning sonar image from the sonar data matrix obtained in step S2; S4, denoising the mechanical scanning sonar image; S5, outputting the denoised image to a target monitoring model for defect identification. The method can not only realize concise and effective denoising and pipeline contour detection, but also realize accurate identification of various defects in municipal pipelines.

The present invention relates to the field of image processing technology, and in particular to a gastric mucosa image analysis method and system for Helicobacter pylori detection. The method comprises the following steps: acquiring gastric mucosa image data; performing image segmentation on the gastric mucosa image data to obtain gastric mucosa image segmentation data, and performing regional characteristic processing based on the gastric mucosa image segmentation data to obtain regional characteristic data; performing image enhancement on the gastric mucosa image data based on the regional characteristic data to obtain gastric mucosa image enhancement data; acquiring region of interest data, and simulating the gastric mucosa image enhancement data based on the region of interest data to obtain region of interest simulation data, so as to perform gastric mucosa image analysis operations. The present invention provides more accurate image analysis data by performing more detailed processing operations on gastric mucosa images.

The invention discloses a mobile detection vehicle suitable for an underground safe operation environment of an underground mine and a using method of the mobile detection vehicle, and relates to the field of mine detection vehicles, the mobile detection vehicle comprises a box body and an avoiding mobile structure installed at the bottom of the box body, the avoiding mobile structure comprises a radar sensor and a mobile motor, and a main controller is installed on one side face in the box body; a radar sensor is installed on the front end face of the bottom of the box body, the top of the box body is open and provided with a matched cover plate, two bottom plates are arranged at the bottom of the box body, bent arms are installed at the two ends of each bottom plate, a moving motor is installed on one end side of each bent arm, and the output end of each moving motor rotationally penetrates through the corresponding bent arm. Real-time detection of front obstacles is achieved, collision accidents are effectively avoided, the safety of underground mine underground operation of the mobile detection vehicle is improved, small ground broken stones or mounds can be avoided, and the adaptability and stability of the mobile detection vehicle are improved.

The application relates to the technical field of vehicles, and provides a front headlamp mounting system and a vehicle. The front headlamp mounting system comprises a front end frame, a support assembly and a front headlamp. The support assembly is connected with the front end frame. The support assembly comprises a plurality of supports. The front headlamp is connected with the support assembly. At least two supports in the plurality of supports are respectively connected to different end portions of the front headlamp, so that the front headlamp is mounted on the front end frame through the support assembly. The above scheme can solve the problem that in the prior art, the front headlamp is mounted on the front bumper area, the front bumper area is crowded in layout, and deviation exists between an actual mounting position and an ideal mounting position of a radar or a camera mounted on the front bumper area.

The present invention discloses a method and system for identifying cracks in a highway asphalt pavement, comprising: obtaining a highway asphalt pavement image; obtaining a highway asphalt pavement crack identification result according to the highway asphalt pavement image and a trained pavement crack identification model, wherein the pavement crack identification model comprises a backbone feature network and a feature fusion network, the backbone feature network comprises a multi-layer feature extraction layer, the multi-layer feature extraction layer continuously extracts multi-layer features from the highway asphalt pavement image, and the features extracted from each layer are input into the feature fusion network, the feature fusion network fuses each layer of features between the first layer and the last layer of features extracted by the backbone feature network with the features of the two adjacent layers to obtain multiple fused features, fuses all fused features to obtain enhanced features, and obtains a highway asphalt pavement crack identification result according to the enhanced features. The accuracy of identifying cracks in a highway asphalt pavement is improved.

The present invention discloses an optical lens, a camera module and an electronic device, wherein the optical lens comprises a first lens having a positive refractive power arranged in sequence from the object side to the image side, wherein the object side surface and the image side surface are convex and concave respectively; a second lens having a refractive power, wherein the image side surface is concave; a third lens having a refractive power; a fourth lens having a refractive power, wherein the object side surface and the image side surface are concave and convex respectively; a fifth lens having a positive refractive power, wherein the object side surface and the image side surface are concave and convex respectively; a sixth lens having a negative refractive power, wherein the object side surface and the image side surface are both concave; and 1.45mm≤Imgh / Fno<1.55mm is satisfied, wherein Imgh is the radius of the maximum effective imaging circle of the optical lens, and Fno is the aperture number of the optical lens. The optical lens, the camera module and the electronic device provided by the present invention can realize the miniaturization of the optical lens, and can also realize the characteristics of large aperture and large image surface, and can improve the imaging clarity in a dark environment.

The utility model discloses an infrared optical system and an infrared optical lens. The infrared optical system sequentially comprises a first aspheric lens with negative focal power, a second aspheric lens with positive focal power, a super lens with positive focal power and a third aspheric lens with positive focal power from an object side to an image side. The surfaces, close to the object side and away from the object side, of the first aspheric lens protrude towards the object side. The surface, close to the object side, of the second aspheric lens protrudes towards the object side, and the surface, away from the object side, of the second aspheric lens protrudes towards the object side or the image side. The surfaces, close to the object side and away from the object side, of the third aspheric lens protrude towards the image side. The infrared optical system provided by the utility model has relatively high resolution and can be matched with an imaging detector with relatively high resolution for use. The infrared optical system has a large imaging range, can obtain more imaging information, and can distinguish more imaging details at the same time. Meanwhile, the infrared optical system also has the advantage of relatively low cost.

The embodiments of the present application relate to a method, device, equipment, medium and program product for train status video recognition, and relate to the field of video information processing technology. The method comprises: performing frame-by-frame inference on each frame sequence image in the target monitoring video through a pre-trained train status recognition model to obtain the state of the vehicle in each frame image of the target monitoring video; if the number of times the same vehicle predicted state appears before the current video frame in the target monitoring video reaches a preset threshold, the vehicle predicted state is determined as the state of the vehicle in the current frame image, and a corresponding video editing is performed according to the current frame image to obtain at least one complete train status video clip corresponding to the target monitoring video. Among them, the train status recognition model is obtained by learning video temporal features based on a 2D convolutional neural network structure. The use of this method can quickly and conveniently identify a complete train video.

An ultrahigh-resolution contrast imaging method, a blood vessel imaging method and an ultrasonic imaging device, the method comprising: acquiring a plurality of frames of ultrasonic data collected for a target object injected with a contrast agent; performing data processing based on the plurality of frames of ultrasonic data using a first processing technique to obtain a first image, wherein the first processing technique is an imaging technique other than a second processing technique; performing data processing based on the plurality of frames of ultrasonic data using the second processing technique to obtain a second image, wherein the second processing technique is an ultrahigh-resolution contrast imaging technique; and displaying the first image and the second image, wherein the spatial resolution of the second image is higher than that of the first image, and the temporal resolution of the second image is higher than or equal to that of the first image. The ultrahigh-resolution contrast imaging method and device of the present application can improve the readability of blood vessel morphology at different levels and more accurately represent the lesion blood vessel morphology and blood perfusion path.