8results about How to "Accurate targeting" patented technology

The invention provides a dexterous manipulator fused with an ontology visual perception function, and belongs to the technical field of electromechanical control. The depth camera is arranged in the palm, the size of the device is remarkably reduced, the size of the device is closer to the size of a human hand, and therefore flexibility and naturalness are improved, the depth camera and a control system are highly integrated, the structure is simplified, synchronism of visual data and hand actions is improved, operation precision is improved, the dexterous manipulator has 16 degrees of freedom, and the flexibility and the reliability of the device are improved. The system is completely driven by a steering engine, the operation flexibility and the fine action execution capability are greatly improved, the requirements of complex tasks are met, the production and maintenance cost is effectively reduced through optimization design and reduction of external accessories, multiple objects are positioned at the same time through a developed instance-level multi-target positioning method and by using a neural network, pre-training is not needed, and the operation efficiency is improved. The space perception capability of the dexterous hand is further improved, and the autonomous operation capability of the dexterous hand in a complex environment is enhanced.

The invention discloses a weak supervision target positioning method based on a semantic-driven Transform model, and the method comprises the following steps: S1, carrying out the preprocessing of a to-be-detected picture, and converting the to-be-detected picture into a format suitable for neural network processing; s2, sending the preprocessed to-be-detected picture and the label of the corresponding picture level into a neural network; the neural network comprises a Transform backbone network, a semantic kernel integration module SKI and a semantic enhancement adaptation module SBA; s3, the neural network processes the picture, and outputs a probability value of the to-be-detected picture corresponding to each category in a training process; s4, outputting the coordinate, category and score of the frame predicted by the to-be-detected picture in the test process; according to the method, the attention graph is more focused on the foreground region, and more accurate semantic perception target positioning is realized in a weak supervision environment, so that the explanation precision and correlation of the model are improved.

The present invention provides a truss robot and its control system applied to an intelligent production line, comprising: a robot body, a main controller, a monocular vision system and a data exchange device; wherein the robot body, the monocular vision system and the data exchange device are electrically connected to the main controller respectively; the monocular vision system collects images of the working area corresponding to the robot body; the main controller inputs the image into a pre-trained and converged neural network model to obtain a recognition result, and controls the robot body action based on the recognition result. The truss robot and its control system applied to an intelligent production line of the present invention recognizes the working area according to the monocular vision system, thereby achieving accurate target positioning, and thereby performing accurate operations, which can cope with the situation of frequent changes in production products and improve production efficiency.

The invention discloses a fully mechanized coal mining face target positioning method based on vision, and the method comprises the following steps: 1, calling two cameras adjacent to a fully mechanized coal mining face target object to obtain a monitoring image, and determining the center coordinates of the target object in the image; 2, the coordinates of the target object are accurately corrected, and the corrected center coordinates of the target object in the monitoring image are obtained; step 3, calculating a horizontal deviation angle of the target in the monitoring image; step 4, based on a binocular ranging principle, measuring and calculating the distance between the target object and the camera device to obtain positioning data of the target object; and step 5, updating the rack number of the adjacent camera at the next moment. According to the method, the horizontal distance between the target object and the camera adjacent to the target object is accurately calculated on the premise that the camera mounting rack number position is known through the target recognition technology and the specific fully-mechanized coal mining face image correction technology, and efficient target positioning of the fully-mechanized coal mining face is achieved.

This invention discloses a pure vision-based 3D reconstruction and adaptive completion method in a collapse environment, belonging to the field of computer vision and emergency rescue 3D reconstruction technology. The method completes initial 3D reconstruction by acquiring data through binocular vision, constructs an anisotropic point cloud density distribution field to achieve accurate detection of low-density blind spots, establishes safe and reachable spatial constraints based on an implicit distance field, and optimizes the completion viewpoint by fusing an observability model. Map iteration is completed through local reconstruction and point cloud updates, and a closed-loop completion mechanism is formed using point cloud density as feedback until the reconstructed model meets the integrity requirements. This invention solves the problems of permanent voids in reconstruction caused by viewpoint blind spots, poor consistency between completion results and the real scene, and insufficient safety in the reconstruction process in existing technologies in collapse environments. It can achieve complete and high-precision autonomous reconstruction of 3D models of collapse disaster sites, providing reliable spatial data support for emergency rescue decision-making.

The present invention relates to the field of signal denoising technology, and in particular to a method for purifying Beidou navigation satellite observation data under a complex electromagnetic background. The method comprises the following steps: obtaining satellite observation site location data and satellite location data; performing satellite signal acquisition on the satellite observation site based on the satellite observation site location data and satellite location data to obtain a satellite observation site receiving signal; discretizing the satellite observation site receiving signal into a multi-angle initial signal to generate an initial satellite observation comprehensive signal discrete graph. The present invention improves the recognition capability and filtering efficiency of interference signals through systematic interference identification, signal purification, multi-angle processing and automated precision optimization, thereby weakening the problem of electromagnetic interference to Beidou navigation satellite signals under a complex electromagnetic background, which not only improves the flexibility of the site selection of Beidou navigation satellite continuous observation stations, but also improves the accuracy and stability of its positioning, navigation and timing services.

The application discloses a kind of intelligent garlic sprout harvesters, it is related to agricultural harvesting machinery technical field, including frame and walking drive mechanism, operation execution mechanism and intelligent sensing and control system installed on frame;The walking drive mechanism includes walking wheel installed on the bottom of frame and driven by motor, realizes self-propelled field operation of whole machine;The operation execution mechanism includes sprout clamping device, needle belt and sprout pulling device sequentially arranged from front to back along the operation direction of frame;The application constructs an intelligent garlic sprout harvesting scheme by machine vision, multi-sensor fusion technology and bionic mechanical structure design, which integrates separating grass, clamping, skinning, pulling and collecting, realizes whole-process automation control from garlic sprout identification, posture righting, skinning loosening, flexible pulling to conveying and collecting, meets the efficient, low-loss, precision harvesting demand in garlic planting scene.

The present invention discloses a target positioning method based on an optoelectronic pod, which belongs to the field of image processing technology and includes the following steps: S1, collecting image data using an optoelectronic pod installed on a drone; S2, generating color array values ​​according to the standard color matrix of the optoelectronic pod and each structural matrix of the image data; S3, determining the target positioning according to the contour of the image data and the color array values. The present invention generates color array values ​​according to the standard color matrix of the optoelectronic pod and each structural matrix of the image data, and the color array values ​​can reflect the color characteristics of the target object, which helps to accurately locate the target in a complex background; the present invention also uses the color array values ​​to analyze the contour of the image data, which can further accurately locate the target.