37results about How to "No loss of detection accuracy" patented technology

The invention discloses an unmanned aerial vehicle tracking method based on an embedded platform. The problem that an occluded vehicle tracked by an unmanned aerial vehicle is lost in the prior art is solved. The method comprises the steps that 1 a vehicle classifier is trained and a map is marked; 2 an image is acquired, and a tracked target is initialized; 3 an image is acquired for target tracking; 4 the classifier is used to determine whether the target is occluded, and if so, the step 5 is carried out, otherwise the step 8 is carried out; 5 whether the target is in an occluded area of the map is determined, and if so, the step 6 is carried out, otherwise the step 7 is carried out; 6 target detection is carried out in the occluded area; the unmanned aerial vehicle hovers; and the step 9 is carried out; 7 the target location is predicted; 8 according to the target, the flight of the unmanned aerial vehicle is controlled; 9 whether the step is end is determined, and if so, tracking is ended, otherwise the step 3 is back. According to the invention, the occluded target can be detected and tracked; the tracking robustness is improved; and the method can be used to track vehicles on urban roads.

The invention discloses a laser in-situ detection system for oxygen concentration in an arsenic-bearing gold concentrate roasting furnace. The system mainly comprises a semiconductor laser, a temperature and current controller for driving the semiconductor laser, a signal generator for generating driving signals, an optical probe for receiving and converting laser beams, a sensor for collecting the temperature and pressure inside the roasting furnace, a data acquisition module for receiving and collecting photoelectric, temperature and pressure signals and an upper computer for computing and detecting the oxygen concentration. The system overcomes the background radiation interference in a combustion environment through the steps of background radiation filtering, baseline correction, absorbance Voigt linear fitting, correction computation of oxygen spectral lines, oxygen concentration computation and inversion and the like, and realizes real-time accurate detection of the oxygen concentration in the roasting furnace.

The invention discloses a method for detecting wireless sensor network sparse events based on compressed sensing and the game theory, which belongs to the field of wireless communication and aims to solve the problem that excessively intensive sensor nodes randomly released in the wireless sensor network probably cause communication interface but sparse sensor nodes probably cause detection leak.The method disclosed by the invention comprises the following steps: (I) dividing M sensor nodes randomly released in the network system into a layers according to the thought of the game theory; (II) determining to activate one or more layers of sensor nodes according to the number (K) of the current events; (III) receiving a data vector y by a data terminal of the network system, and obtaining a measurement matrix [phi]P*N according to the detection process; and (IV) obtaining the specific distribution vector E of K happening events in N event sources by use of a greedy sparse event detection algorithm.

The invention discloses a method and equipment for simultaneously inspecting a plurality of liquid substances by using dual-energy computed tomography (CT). The method comprises the following steps of: performing dual-energy CT scanning on inspected liquid substances at a plurality of positions to acquire dual-energy CT protection data; performing CT reconstruction according to the projection data to acquire the CT images for representing physical attribute values of the inspected liquid substances; rotating the CT images of the inspected liquid substances after performing the dual-energy CT scanning at each position so as to align the images formed by multiple times of scanning operation; extracting the physical attribute values of the inspected liquid substances according to the CT images; and judging whether the inspected liquid substances are dangerous substances based on the physical attribute values. By the method, the liquid substances can be inspected rapidly.

The objective of the present invention is to provide an illuminating device which uses one illuminating light source to supply an amount of light onto an object being detected, to create a detection image which is bright overall and which has no parts in shadow. A light source required to illuminate the object being detected from above and below is shared, and the object being detected is illuminated by the light from the illuminating light source being received by a plurality of reflectors and divided to form upward illuminating light and downward illuminating light. Further, optical paths ofat least some light beams from the illuminating light source are reflected by a refracting mirror and are then caused to be incident on the reflectors. In this way, the number of illuminating light sources can be reduced without causing a reduction in the amount of light, and a reduction in the size of the overall illuminating device can be achieved. Furthermore, by employing a concave curved mirror reflector consisting of a plurality of concave curved partial surfaces having different curvatures, the object being detected is present in an illuminated zone even if a fall trajectory of the object being detected deviates from an assumed fall trajectory, and therefore a shadow does not occur in a detection image obtained using a camera, and the image itself does not become dark. As a result,a deterioration in detection accuracy can be prevented.

The invention discloses a method and equipment for checking a liquid object by dual-energy computed tomography (CT). The method comprises the following steps of: performing data register (DR) scanning on the liquid object to acquire the transmission image of the checked liquid object; analyzing the pixel value of the transmission image to judge whether the transmission image is layered or not; performing dual-energy CT scanning on the checked liquid object at the at least one position of each determined layer to acquire dual-energy CT projection data; performing CT reconstruction according tothe projection data to acquire a CT image which indicates the physical attribute value of the checked liquid object; exacting the physical attribute value of the checked liquid object according to the CT image; and judging whether the checked liquid object is an dangerous object or not based on the physical attribute value. By the method, the liquid product can be quickly checked.

The invention discloses a method and equipment for simultaneously inspecting a plurality of liquid substances by using dual-energy computed tomography (CT). The method comprises the following steps of: performing dual-energy CT scanning on inspected liquid substances at a plurality of positions to acquire dual-energy CT protection data; rotating CT images of the inspected liquid substances after performing the dual-energy CT scanning at each position so as to align the images formed by multiple times of scanning operation; performing CT reconstruction according to the projection data to acquire the CT images for representing physical attribute values of the inspected liquid substances; extracting the physical attribute values of the inspected liquid substances according to the CT images; and judging whether the inspected liquid substances are suspicious based on the physical attribute values and reference physical attribute values of the liquid substances. By the method, the liquid substances can be inspected rapidly.

The present invention provides a face detection optimization method based on a deep convolutional cascaded network, which specifically includes the following steps: using the deep cascaded network to detect areas where human faces may appear, that is, hot spots; updating the hot spots, Set all the areas not in the hot zone to zero; perform data sparse compression on the obtained whole image. The invention greatly reduces the amount of redundant computation and improves the operating efficiency of the algorithm. The efficiency of the CNN network face detection algorithm on the front-end embedded platform can be increased by 20-30%, and the method of the present invention will not lose the detection accuracy, providing a better solution for the faster operation of the deep convolutional network on the front-end The scheme laid the foundation for the large-scale application of the network in the later stage.

The invention discloses a method and equipment for confirming a CT (Computed Tomography) scanning position in a dangerous goods inspection system. The method comprises the following steps of: firstly, through a CT scanning and dual-energy rebuilding method, obtaining one or more layers of CT images containing physical attributes of liquid to be tested; secondly, through an image processing and analyzing method, obtaining a physical attribute value of each liquid to be tested from the CT images; and finally, judging whether the liquid-state object to be tested is a dangerous goods or not according to the physical attribute value. The CT scanning method comprises a regular sectional CT scanning technology and a spiral CT scanning technology; and when the regular sectional CT scanning technology is utilized, a series of specific positions can be set so as to be scanned, a scanning position can be specified by an operator through a DR (Digital Radiography) image, and the position of a liquid part can be confirmed through automatic analysis on the DR image, so that CT scanning is guided.

The invention discloses a vehicle tracking method for an unmanned aerial vehicle based on an embedded platform, which mainly solves the problem in the prior art that the unmanned aerial vehicle loses tracking when vehicle tracking is blocked. The implementation steps are: 1. Train the vehicle classifier and mark the map; 2. Get a frame of image and initialize the tracking target; 3. Get a frame of image for target tracking; 4. Use the classifier to judge whether the target is blocked, if so, Go to step 5, otherwise go to step 8; 5. Determine whether the target is in the occluded area of ​​the map, if yes, go to step 6, otherwise go to step 7; 6. Perform target detection in the occluded area and make the drone hover, go to step 9; 7. Predict the target position; 8. Control the flight of the UAV according to the target; 9. Determine whether it is over, if so, end the tracking, otherwise return to step 3. The invention can detect and track the blocked target, improves the tracking robustness, and can be used for tracking vehicles on urban roads.