31results about How to "Improve detection and positioning accuracy" patented technology

The invention proposes an object detection and location method and system based on depth data. The method comprises the steps of acquiring a depth map by using depth data acquired by a depth data sensor, constructing a background model of the depth map, acquiring an object foreground region under the perspective of the depth map, and carrying out initial segmentation of the object foreground region; mapping the object foreground region after the perspective of the depth map to an overhead perspective by using a projection matrix, and segmenting the object foreground region under the overhead perspective again according to the location difference of an object in the horizontal space; extracting depth features of the foreground region, training an object detection module, and carrying out detection positioning for adhered objects by using the object detection module; and carrying out three-dimensional coordinate location correlation tracking for detection positioning results on an overhead map, mapping the detection positioning results to an original depth map coordinate, and obtaining a final object detection result. The invention can complete accurate object detection and location in a quick and robust manner under severely blocked and changing illumination conditions.

The invention discloses a method for recognizing an object in an image based on depth learning, Here are the steps: an image is inputted, the candidate regions are extracted by convolution neural network, filter and optimize the output candidate region, At the same time, each candidate region is normalized, the candidate region is input to convolution neural network for feature extraction, and thetrained classification and regression network is used to classify, locate and detect the target image. Finally, the selected target region is subjected to border regression operation to correct the position of the target region. This method uses convolution neural network to extract the regions which may contain objects in the image, and reduces the number of candidate regions. At the same time,it optimizes the filtering operation on the output candidate regions of convolution neural network, which improves the computational speed of the algorithm. In addition, a variety of length-width ratio and region size are used for the candidate region of the target detection, which is closer to the real scene and improves the robustness of the algorithm.

The invention provides a dynamic balance test method and a device thereof. The dynamic balance test method is used for testing workpieces with peripheries distributed with grooves uniformly and includes steps as follows: axially rotating a workpiece; utilizing a non-contact sensor to test the grooves on the workpiece at a fixed position and generate a pulse signal, wherein the pulse signal generated by the non-contact sensor when the workpiece begins to rotate is used as an unbalance datum zero-degree signal; dividing frequency of the pulse signal, wherein multiple of frequency dividing is equal to the amount of the grooves on the workpiece; and defining unbalance amount and unbalance angle according to the pulse signal after frequency dividing, wherein a first pulse front edge of the unbalance datum zero-degree signal is used as a zero-angle position of the workpiece when defining the unbalance angle. The dynamic balance test method and the device thereof remove special processes on the surface of the workpiece and save working procedures, thereby increasing positioning accuracy of unbalance angle tests. The dynamic balance test method and the device thereof are adoptable to automation of dynamic balance.

The invention discloses a passive acoustic detection positioning device for a low-altitude unmanned plane, and the device comprises an acoustic sensor array detection module, a multichannel data collection module, a signal preprocessing module, an Ethernet transmission module, an embedded processing module, a display and control module, and a voice alarm module. The acoustic sensor array detectionmodule is used for receiving an acoustic signal, and the multichannel data collection module is used for converting the collected acoustic signal into a digital signal. The signal preprocessing module is used for the filtering and denoising of the signal, and transmitting the processed signal to the embedded processing module through the Ethernet transmission module, so as to achieve the signal processing and upper computer information processing. The display and control module is used for displaying detection information and sending a related instruction to an alarm and countermeasure signal. The voice alarm module is used for the real-time voice broadcasting of the detection and positioning information. The device is high in detection and positioning precision, can achieve the real-timedetection and display of the positions, numbers and types of consumption-level slow and small flying objects invading an airspace, and improves the low-altitude early warning defense strength.

An AI and vision combined intelligent container truck positioning method and device calibrate a vehicle parking point through a camera in advance, and the AI and vision combined intelligent containertruck positioning method comprises the following steps: 1) reading a video stream of the camera, decoding the video stream into a picture, and preprocessing the picture; 2) inputting the processed picture into a pre-trained Mask RCNN neural network for deep learning for positioning to obtain a central point of the container or the frame, calculating an Euclidean distance between the central pointand a calibrated vehicle parking point, if the Euclidean distance is greater than a preset value, returning to the step 1), and otherwise, entering a step 3); and 3) matching the picture with a template of the container or the frame by adopting a template matching algorithm, and outputting an optimal solution to obtain the position of the container or the frame.

The invention discloses an underground pipeline detection device. The objectives of the invention are to improve underground pipeline detection positioning precision and acquire specific distributioninformation of underground pipelines. The device comprises a support, a controller, at least two signal emitters and at least two signal receivers, wherein the controller, the signal emitters and thesignal receivers are installed on the support. The signal emitters and the signal receivers are connected with the controller. The signal receivers are used for receiving signals emitted by the signalemitters. The controller is used for extracting signals reflected by the underground pipelines from the signals received by the signal receivers, and acquiring position information of the undergroundpipelines according to the reflected signals.

The invention discloses a container truck head anti-smashing protection system and method, a computer storage medium and a gantry crane. The system comprises a calibration module for carrying out calibration on a parking point of a container truck in advance; an acquisition module for acquiring the video stream of the calibration module, decoding the video stream into a picture and preprocessing the picture; an AI visual algorithm module for inputting the preprocessed picture into a Mask RCNN neural network algorithm model to obtain a truck head instance segmentation graph and truck head center point coordinates of the container truck; a judgment module for obtaining a parking point of a calibration container truck of the calibration module and a truck head center point of the AI visual algorithm module and calculating the distance between the parking point of the calibration container truck and the truck head center point; and a data recording module for recording historical records of each module. According to the invention, real-time detection, identification and positioning tracking of the container truck head are realized.

The invention relates to an intelligent household indoor air quality detection and emergency oxygen supply device. The intelligent household indoor air quality detection and emergency oxygen supply device comprises a bearing base, a reaction cavity, electrolysis electrodes, electric heating wires, an oxygen exhausting pipe, a hydrogen exhausting pipe, an air quality detection device and a controlcircuit, wherein the reaction cavity, the air quality detection device and the control circuit are all installed on the upper surface of the bearing base; the control circuit is electrically connectedwith the reaction cavity, the electrolysis electrodes, the electric heating wires and the air quality detection device; the reaction cavity is of a closed cavity structure with the axial cross section being square, and comprises a bearing groove and a sealing cover; the oxygen exhausting pipe and the hydrogen exhausting pipe are both connected with the reaction cavity through exhaust ports; the electrolysis electrodes are installed at the bottom of the bearing groove; and the electric heating wires are uniformly distributed on the inner side face of the bearing groove. According to the intelligent household indoor air quality detection and emergency oxygen supply device, on one hand, the indoor air quality can be effectively detected, on the other hand, when the indoor oxygen content is low, oxygen is supplemented for the indoor air, and the requirement of assisting oxygen uptake can also be directly provided for a user.

The invention discloses a precise facial paralysis degree evaluation method and device based on semantic segmentation. The method comprises the following steps: establishing a facial paralysis semantic segmentation model; acquiring to-be-detected data and processing the to-be-detected data; sequentially inputting the expression-free natural state static image, the sequence image I, the sequence image II, the sequence image III and the sequence image IV into the facial paralysis semantic segmentation model to output a plurality of corresponding groups of face shapes, and updating the pluralityof groups of face shapes; evaluating the facial paralysis degree of the user to be detected; calculating theta1, theta2, theta3, theta4, theta5, theta6, theta7, theta8, theta9, theta10, b1, b2, c1, c2, e1 and e2, and comparing the calculated values with the threshold values; and judging the facial paralysis degree of the to-be-detected user through the comparison result, and calculating a facial paralysis index. According to the invention, the detection model has high detection positioning precision, the precision and accuracy of comprehensive evaluation and detection of the facial paralysis degree of the to-be-detected user are greatly improved, and a powerful support is provided for prevention, discovery and treatment of facial paralysis patients.

The invention provides an instrument detection method based on a one-shot mechanism. The method comprises the following steps: firstly, performing enhancement processing on a collected instrument image data set by utilizing size transformation and perspective transformation; constructing an instrument detection network model based on a one-shot mechanism, wherein the instrument detection network model comprises a feature extraction twin network module based on ResNet-18, an RPN module, a ROIAlign pooling layer and a feature fusion network module based on Grid Head exchange; training the network model by using the enhanced image data set to obtain a trained network; and finally, processing a to-be-detected instrument image by using the trained network to obtain a final detection result. The method has the capability of detecting the dial position of any quadrilateral instrument, and has relatively high instrument detection precision.

The invention provides an anti-intrusion artificial intelligence radar video monitoring system, and relates to the technical field of artificial intelligence and radar technology application, and the video monitoring system is mainly composed of a radar device, a radar information processing workstation, a video monitoring camera, a data processing server, a communication link device and a terminal device. A radar positioning technology, a video monitoring technology and an artificial intelligence visual technology are effectively combined and applied to regional intrusion prevention and control, the type of an intrusion target can be actively identified, the intrusion target can be positioned and the type of the intrusion target can be identified without matching with vehicle-mounted or airborne response equipment, and judgment and taking of countermeasures are facilitated; the video monitoring system provided by the invention can display the video image of the intruding target in real time, so that the intruding target can be eliminated in time.