226 results about "Color detection" patented technology

A recommendation system identifies resale goods corresponding to retail goods being viewed by a user on a retail goods website. Retail goods metadata and images are extracted from the retail goods website using heuristics and LLM-based methods. The extracted retail goods data undergoes ML model-based product category classification, intelligent image cropping, and color detection. Vector embeddings are generated for images and text using ML models and compared to resale goods vector embeddings stored in a vector database. Multiple result sets are retrieved, fused, and re-ranked using LLM-based and preference-aware re-ranking. A data pipeline continuously loads, cleans, and processes resale goods inventory from resale goods websites.

The invention relates to the technical field of image recognition, in particular to an agricultural crop disease and pest detection method and system based on image segmentation, and the method comprises the steps: collecting a vertical view angle image and an inclined view angle image of a tea garden through an unmanned aerial vehicle, carrying out the image registration through SIFT feature extraction, KNN matching and RANSAC filtering, and carrying out the alignment to a reference image coordinate system. And segmenting the foreground region by adopting a U-Net algorithm, and performing post-processing optimization. Positioning a disease area through a color detection method, extracting disease features, and identifying a disease stage by using a CNN algorithm; and using a YOLOv3 algorithm to identify insect body areas, extracting insect pest features, and evaluating insect pest degrees. And finally, based on the disease stage and the pest degree, using an LSTM algorithm to predict the pest diffusion trend, and generating a space thermodynamic diagram according to the reference image coordinate system, thereby improving the detection precision and timeliness, and providing technical support for intelligent agricultural prevention and control.

Embodiments of the present application relate to the technical field of image recognition, and disclose a color band resistor color extraction method and apparatus, a device, and a medium, which solve the technical problem in the prior art of poor precision of color band resistor color detection. The color band resistor color extraction method comprises: acquiring an RGB color image of a target color band resistor; acquiring an LAB color image on the basis of the RGB color image; performing transformation processing on an LAB value corresponding to each color type of the color band resistor to acquire a color extraction threshold; and performing color extraction processing on the LAB color image on the basis of the color extraction threshold. The method of the present application effectively improves the precision of color band resistor color detection.

The invention discloses a display light color detection system, and relates to the technical field of low-brightness display light color detection, and the system comprises a dynamic modulation module which is configured to carry out the high-frequency pulse modulation of a dark-state signal of a detection object, and is used for enabling the dark-state signal to carry a modulation component of a characteristic frequency; the time sequence division acquisition unit comprises a high-speed optical shutter assembly and a detector, and the high-speed optical shutter assembly synchronously switches the on-off state of an optical path according to the characteristic frequency; according to the invention, through high-frequency pulse modulation of dark state signals, time-sequence-divided acquisition of noise and signals and time-frequency domain combined decoupling, traditional hardware limitation is broken through, ultra-low brightness signals and stray light interference are effectively separated, 0.01 cd / m < 2 >-level high-precision detection is realized, the detection efficiency is improved by more than 7 times, and the misjudgment rate is significantly reduced. The problems of missing detection of a novel display technology and the yield of an automatic production line are solved, and an active coding-time sequence calibration-cross-domain cooperation framework has remarkable innovativeness and industrial value.

The invention relates to the technical field of color detection, in particular to a color difference calculation optimization method and system for archaized linen viscose reactive dyeing. Comprising the following steps: monitoring a fabric motion state in real time, calculating a motion speed and displacement, and determining a spectrum acquisition opportunity and a position compensation parameter; illuminating the surface of the to-be-detected fabric at an acquisition position through a polarized light source, and performing polarized spectrum data acquisition in the normal direction of the surface of the fabric to obtain reflection spectrum information; exciting the surface of the to-be-detected fabric by adopting an ultraviolet light source, collecting fluorescence spectrum data, and identifying different fiber areas; calculating fiber density distribution in the fiber area, and constructing a layered weight system; calculating chromatic aberration of the fiber area based on the reflection spectrum information, extracting chromatic aberration characteristics, and measuring and calculating an antique degree quantification result; setting a color difference threshold value standard according to an antique degree quantification result, and adjusting a color difference threshold value according to an evolution trend; and based on the layered weight system and the chromatic aberration threshold, performing weighted calculation on the chromatic aberration value of the fiber region to obtain a comprehensive chromatic aberration evaluation result.

The utility model relates to the technical field of chilli fragment color detection, and discloses a color detection device for chilli fragment production, which comprises a detection frame and a production frame, a driving mechanism is fixedly arranged on the inner side of the detection frame, a transparent screening mechanism is fixedly arranged at the output end of the driving mechanism, the driving mechanism comprises a connecting line, and the connecting line is connected with the production frame. An output rotating shaft is fixedly arranged on one side of the connecting line, a cross-shaped driving rod is fixedly arranged on a shaft body of the output rotating shaft, the transparent screening mechanism comprises a transparent cover, L-shaped fixing plates are fixedly arranged on the periphery of the inner side of the transparent cover, and the side, away from the output rotating shaft, of a rod body of the cross-shaped driving rod is connected to the inner side of the transparent cover. The picking mechanism guides a part of crushed chilies in the production process, the crushed chilies are continuously turned over through the transparent screening mechanism so that a chromatographic instrument can conveniently detect the crushed chilies, the sampling time of production personnel and the time for turning over the crushed chilies in the chromatographic instrument are shortened, and the production efficiency is improved.

Disclosed is a surveillance camera capable of detecting a color of a vehicle using vehicle logo detection. This specification describes a method that, instead of processing an entire image of a vehicle, detects the vehicle's logo and expands a bounding box of the detected logo to analyze a dominant color within the expanded bounding box for higher accuracy. During a dominant color analysis process, color correction may be performed based on metadata, environmental factors, and a lighting condition. In addition, more accurate vehicle color detection may be achieved through color distance measurement. According to this specification, at least one of a surveillance camera, an autonomous vehicle, a user terminal, and a server may be associated with an Artificial Intelligence (AI) module, a robot, an Augmented Reality (AR) device, a Virtual Reality (VR) device, and a 5G service.

The invention discloses an intelligent monitoring system for camellia oleifera processing based on artificial intelligence, and relates to the technical field of camellia oleifera fruit processing, and the system comprises a temperature monitoring module which comprises a multi-point temperature sensing array which is arranged along the axial direction of a presser and is installed at the front end, the middle part and an oil outlet of a barrel, the temperature sensor is used for collecting temperature and forming temperature distribution data in the barrel; the color detection module is arranged at an outlet of the oil outlet pipeline, obtains oil chromaticity parameters and a yellow-red ratio through spectral imaging, and extracts color feature data; and the volatile matter detection module is arranged above the discharged oil collection cavity and is used for detecting the concentration of volatile organic compounds through a gas sensing array and generating flavor characteristic data. By arranging the temperature monitoring module, the color detection module and the volatile matter detection module, real-time collection of multi-dimensional information in the squeezing process is achieved, and based on an asynchronous triggering and weighted fusion mechanism, space-time alignment and conjoint analysis are conducted, a joint feature vector is constructed, and a real-time prediction result is output.

The invention relates to the technical field of detection devices, in particular to a detection device for rapidly measuring the specific gravity, turbidity and color of urine, which comprises a mounting assembly, a through prism, a specific gravity detection unit, a turbidity and color detection unit, a color detection unit and a temperature detection unit, light rays emitted by the specific gravity detection unit enter from one side of the through prism to be totally reflected, receive and collect data, pass through an established fitting function and output a specific gravity value, light rays emitted by the turbidity and color detection unit penetrate through a sample, receive data of scattered light rays passing through the sample, obtain turbidity transmission data and color transmission data, and output a turbidity value in combination with the fitting function. The color detection unit emits light to penetrate through the sample, receives light data reflected by the sample, obtains color reflection data, combines the color reflection data with the color transmission data to establish a fitting function, and outputs a color value, the detection device uses less sample detection, integrates urine specific gravity, turbidity and color detection, is high in automation degree, and improves the detection accuracy.

The invention relates to the technical field of machine vision detection, and discloses a color printed matter quality detection method which comprises the following steps: acquiring original light field response data of the surface of a printed matter in a preset structure light field; carrying out decoupling calculation on the data to generate a basic color graph representing intrinsic colors, a glossiness graph representing process characteristics and a surface normal graph representing geometrical morphology; the method is characterized in that a texture area is identified based on a surface normal graph of a standard sample, and a dynamic reference model is established for the area in combination with a basic color graph so as to provide a self-adaptive color detection threshold value; and finally, detecting color, structure and process flaws in parallel by utilizing each physical characteristic spectrum and the dynamic reference model. According to the method, through decoupling and information guidance of physical dimensions, interference of physical textures on color detection can be fundamentally inhibited, the false alarm rate is remarkably reduced, meanwhile, multi-dimensional comprehensive evaluation of flaws is achieved, and the detection result is objective and reliable.

The embodiment of the invention provides a dust display system and method based on polychromatic light and autonomous cleaning equipment, and belongs to the field of intelligent cleaning equipment. The system comprises a multi-color detection light emitter which is configured to project at least two detection light beams with different wavelengths to a to-be-cleaned ground area in front of or laterally in front of the autonomous cleaning equipment; the light receiver is configured to receive a light signal from the ground area to be cleaned and generate corresponding image data; the control module is electrically connected with the multi-color detection light emitter and the light receiver, and is configured to control the multi-color detection light emitter to switch and emit detection light beams with different wavelengths in a time-sharing mode; and on the basis of multiple groups of image data acquired by the light receiver under the irradiation of the detection light beams with different wavelengths, smudginess identification is performed, and a cleaning strategy control signal is output. The ground background and real dirt can be effectively distinguished, and the false alarm rate is greatly reduced; and meanwhile, the identification capability of smudginess colors and properties is improved, and the method is suitable for diversified family smudginess scenes.

The present invention relates to the technical field of image processing, and in particular to a method and system for analyzing the flame of a hydrogen chloride synthesis furnace, including using a flame sensor to determine whether there is a flame in the hydrogen chloride synthesis furnace; collecting a flame video and obtaining a flame image by frame division; performing color detection on the obtained image under the RGB and HSI mixed color models to obtain a flame pixel point distribution matrix and obtain a preliminary image containing a suspected flame area; performing grayscale processing on the preliminary image containing the suspected flame area, using PSO search for global optimization, and substituting the global optimal solution back into the K-means clustering algorithm to perform clustering on the image; multiplying the matrix of the preliminary image containing the suspected flame area by the matrix of the flame image after clustering to obtain the final flame area; adjusting the gas concentration according to the RGB mean value and color temperature of the final flame area. The present invention solves the problem of poor control quality of manually observing the flame color.

The utility model belongs to the technical field of medical detection, and provides a urine analyzer. The urine color turbidity specific gravity conductivity detection device comprises a sample suction device, a sample conveying pipeline and a urine color turbidity specific gravity conductivity detection device, and the urine color turbidity specific gravity conductivity detection device comprises a shell and a prism module arranged on the shell. The turbidity detection module, the specific gravity detection module, the color detection module and the conductivity detection module are arranged on the prism module, the prism module is provided with a sample channel, and the turbidity detection module, the specific gravity detection module and the color detection module are located in the circumferential direction of the sample channel. The conductivity detection module is positioned at the end part of the sample channel; the two ends of the conveying pipeline are communicated with the sample suction device and the sample channel respectively, so that the sample sucked by the sample suction device is conveyed into the sample channel. The utility model is favorable for reducing the use amount of samples.

The invention relates to an intelligent color box printing color analysis control method, and provides a reinforcement learning decision closed-loop control method integrating distributed color detection, space-time multi-modal data processing, feature extraction and self-supervised clustering and multi-scale reward modeling. Multi-variable normalization and synchronization are completed through multi-point real-time collection of color, process and environment data, color change meta-state features are extracted based on a self-encoding neural network, and a meta-state distribution model related to space and working conditions is established through self-supervised clustering. And the color risk is evaluated in combination with a multi-scale reward function, so that color parameter optimal adaptive adjustment and cooperative control driven by reinforcement learning are realized. In an actual control link, the system can dynamically optimize a meta-state aggregation and strategy network structure based on a feedback signal, a stable and self-adaptive color consistency closed-loop control system is constructed, and the color consistency and process intelligence level of a printed product under a complex working condition are effectively improved.

The invention discloses a traffic light state detection method, apparatus and device, and a computer program product. The traffic light state detection method comprises the steps of obtaining an image of a traffic scene at a current moment; traffic light detection is carried out on the image of the traffic scene at the current moment, a traffic light detection result at the current moment is obtained, and the traffic light detection result at the current moment comprises an initial traffic light color detection result at the current moment; and determining an actual traffic light state at the current moment by using a predefined finite state machine and a preset smoothing mechanism according to the initial traffic light color detection result at the current moment. According to the traffic light state detection method provided by the embodiment of the invention, the traffic light state is identified by using the finite-state machine and the smoothing mechanism of the multi-frame data, which is different from a scheme which only depends on single-frame detection or a single decision model, so that the accuracy and robustness of traffic light state identification are remarkably improved; and identification challenges in a complex traffic environment can be effectively handled.

The present invention relates to a method and a dual-color detection kit for the rapid and simultaneous detection of MS-222 and malachite green, and a technology for the rapid detection of veterinary drug residues in aquatic products. The invention utilizes red and blue colloidal gold with different particle sizes, combined with specific antibodies to construct different red and blue probes, respectively. The invention utilizes the principle that MG-BSA complete antigen and 3-aminobenzoic acid-BSA antigen coated on the T1 line and the T2 line, respectively, can compete with MG and MS-222 in the sample for binding to the blue probe and the red probe, thereby constructing a dual-color rapid detection test strip for MS-222 and malachite green that is clearly distinguishable to the naked eye and simple to operate.

The invention relates to the technical field of hotpot condiment stir-frying, in particular to a caviar hotpot condiment preparation method and equipment thereof.The caviar hotpot condiment preparation method comprises a pot body, a condiment stirring mechanism, a heater, a refrigerator and a quality detection assembly, and the quality detection assembly comprises a supporting plate, a telescopic sampling pipe, a dilution bin and a full-automatic Lovibond colorimeter; a telescopic sampling pipe extends into the pot body, mixed grease of a base material is pumped into a dilution bin, then soybean oil is added into the dilution bin for ten-time dilution, the diluted mixed grease is pumped into a full-automatic Lovibond colorimeter for color detection, and if the color value of Lovibond is within the range of 7.0-8.5, it is indicated that the color is qualified; therefore, manual sampling is not needed, the hotpot condiment can be automatically diluted, the detection efficiency is remarkably improved, color and luster detection is performed through the full-automatic Lovibond colorimeter, the accuracy can be remarkably improved compared with visual inspection, and the color and luster quality of the hotpot condiment is guaranteed.

A system and method is presented for hands-free operation of elevator using a smart phone or like device to display one or more colours that are read by a colour-detection device mounted in an elevator cabin and at each elevator landing. Different colours can be displayed by the smart phone or like device and detected by the colour-detection device where different can represent different commands to operate the elevator.

The invention discloses a fruit quality comprehensive detection method based on machine vision and X-rays. The fruit quality comprehensive detection method is used for performing graded detection on appearance, internal defects and sugar degree of fruits. The appearance detection comprises the steps of constructing a fruit surface defect feature matrix, a fruit shape feature matrix and a color feature matrix, and inputting the fruit surface defect feature matrix into a BPNN network to realize fruit surface defect detection; inputting the fruit shape feature matrix into another BPNN network to realize fruit shape detection; inputting the color feature matrix into an SVM model to realize color detection; the internal defect detection comprises the steps of constructing a plurality of different internal defect detection models, and fusing results of the internal defect detection models based on a multi-criterion decision to obtain an internal defect detection result; the sugar degree detection comprises the steps of extracting a pulp area based on X rays, extracting morphological characteristics and gray statistical characteristics, predicting the sugar degree by utilizing a sugar degree prediction model, and grading the sugar degree according to a sugar degree prediction value. The method can more comprehensively reflect the comprehensive quality of the fruits.

The invention discloses textile color detection equipment which comprises a base, two sets of supporting rods are symmetrically and fixedly arranged on the base, side plates are fixedly arranged on the two sets of supporting rods, two rotating rollers are rotationally arranged between the two side plates, and the two rotating rollers are symmetrically and rotationally arranged at the two ends of the two side plates through shaft rods. The two rotating rollers are sleeved with a conveying belt, a plurality of sets of clamping plates are evenly arranged on the conveying belt, each set of clamping plates comprises two clamping plates symmetrically arranged on the two sides of the conveying belt, each locking bolt is rotationally arranged on the conveying belt, and each clamping plate is provided with a threaded through hole corresponding to the corresponding locking bolt. And a detection chamber is fixedly arranged on the two side plates. The textile can be smoothed before being detected, the situation that the color detection accuracy is affected by shadow generated when the textile is folded is avoided, the detection lens can be quickly wiped when needed, and the color detection quality is guaranteed.

The invention provides a portrait picture processing method and apparatus, and an electronic device. The method comprises the steps of obtaining an uploaded portrait picture; performing clothing feature detection and clothing color detection on the portrait picture to determine a first clothing feature and a clothing color; performing feature matching from a preset clothing asset library according to the first clothing feature, and determining a target clothing asset; and according to the garment color, performing graded dyeing on the target garment assets to obtain virtual role garments. According to the method, the technical capabilities of game clothing asset retrieval and clothing asset grading dyeing can be introduced into an intelligent image creation function, and the problem that in an existing intelligent image creation function, a pinched virtual character image is different from an overall image of an uploaded portrait picture is solved.

The invention discloses an intelligent laser system with an energy-saving function. The intelligent laser system comprises a color detection device, a height measurement device, a laser generation device, a temperature detection device, a conveying device and a controller. According to the device, edge sealing belts with different colors can be detected through the color detection device, edge sealing belts with different heights are measured through the measuring device, and meanwhile, the controller controls the output power of the laser generating device and the size of light spots in real time according to the conveying speed of the conveying device; when the laser generating device outputs laser to process the edge sealing band, the controller detects the temperature of the laser irradiation area through the temperature detection device and adjusts the output power and the light spot size of the laser generating device to form closed-loop control, so that the edge sealing bands with different colors and heights can be processed by proper output power and light spots, and the processing efficiency of the edge sealing band is improved. And a proper processing temperature is achieved, so that higher processing quality is obtained.

The invention relates to the technical field of image processing, and discloses a vision-based bad part detection method and device. The method comprises the steps of collecting a target image of a region of interest, wherein the region of interest comprises a standard region where a to-be-detected part is located; detecting the target image based on a part detection model, and determining whether the to-be-detected part belongs to a bad part or not; wherein the part detection model comprises at least one of a part missing detection model, a part offset detection model and a part defect detection model; the part missing detection model detects whether a to-be-detected part is missing or not based on the color of the to-be-detected part, and the part offset detection model is used for detecting whether the to-be-detected part is offset or not based on the part characteristics of the to-be-detected part. The part defect detection model is used for detecting whether the part to be detected has defects or not based on a deep learning model. The detection efficiency, the detection precision and the detection time of defective parts can be improved.

The utility model provides a quick screening device for red shell clams, which is characterized by comprising a conveying component, a color detection component and a screening mechanism, the conveying component comprises an inlet bin and a first guide groove, an opening of the inlet bin is arranged at the top end of the device and is communicated with the first guide groove, and the color detection component is arranged in the first guide groove. Limiting clamps are symmetrically arranged at an inlet of the first guide groove, the color detection assembly comprises color sensors, a data processor and a second guide groove, an inlet of the second guide groove is communicated with an outlet of the first guide groove, and the color sensors are arranged in the second guide groove and are in signal connection with the data processor; the screening mechanism comprises a movable push plate and a temporary storage bin, a hole is formed in the movable push plate, a sliding way is arranged below the hole, and the sliding way communicates with the temporary storage bin. According to the utility model, the problems of low efficiency and low screening accuracy of the existing manual screening technology are solved.

The invention discloses a tooth detection method and device, and relates to the technical field of artificial intelligence and medical health. A specific embodiment of the method comprises the following steps: recognizing face key points according to an obtained to-be-detected image; in response to the fact that the identification result does not meet the preset detection condition, recursively extracting mouth features from the to-be-detected image, and determining a mouth region based on the mouth features; and identifying a tooth area from the mouth area, and performing color detection processing on the tooth area to generate a tooth detection result. According to the embodiment, double mechanisms of face key point detection and the feature classifier are fused, face key point recognition is preferentially carried out, and when the main detection fails, the feature classifier is switched to carry out standby detection, so that the tooth detection efficiency and the accuracy of a detection result are improved, and the user experience is improved.

The invention relates to a method for detecting the peroxide value and the acid value of food, and belongs to the technical field of detection instruments. Putting a to-be-detected sample into the reagent bottle, and then dropwise adding a detection reagent into the reagent bottle; in the process, a light source is used for irradiating the reagent bottle, and a sensor is used for receiving light reflected by the reagent in the reagent bottle; the control system detects the light received by the sensor; when the color of the received light is a preset value, dropwise adding of the detection reagent is stopped; recording a titration end point, and determining a peroxide value; the acid value can be measured by the same method, and the color of the solution judged by the titration end point for measuring the peroxide value is different. According to the invention, light is utilized to irradiate a reagent and a sensor is utilized to carry out color detection, so that automatic detection of a peroxide value and an acid value can be realized, the detection precision can be improved, automatic detection can be carried out for a long time, and the detection efficiency is improved; meanwhile, workers are prevented from inhaling gas with pungent smell, and harm of chemical reagents to experimenters is reduced.

To prevent deterioration in measurement accuracy due to the influence of stray light in a technique for inspecting an image recorded on a print medium by using an imaging part and a colorimetric part.SOLUTION: In the image inspection apparatus and the printing apparatus according to the aspects of the invention, the printing medium is supported by being stretched between the first roller and the second roller. An imaging unit that images an imaging region by causing illumination light to be incident on a portion of the printing medium wound around the first roller and receiving reflected light from the imaging region on which the illumination light is incident, a colorimetry unit that performs color detection of a color detection region by causing the illumination light to be incident on a portion of the printing medium wound around the second roller and receiving light reflected from the color detection region on which the illumination light is incident and incident on a light receiver, and a light shielding unit that is disposed between the imaging unit and the light receiver on a first main surface side of the printing medium stretched between the first roller and the second roller and shields light from the imaging unit toward the light receiver.SELECTED DRAWING: Figure 7

The utility model provides an identification circuit and an electronic device, and the identification circuit comprises a light detection circuit which is used for receiving a light intensity signal corresponding to a light source signal and outputting a voltage signal corresponding to the light intensity signal; and one end of the color identification circuit is connected with the other end of the light detection circuit, and the color identification circuit is used for receiving the voltage signal and outputting a color identification result. According to the utility model, the color of the electronic equipment shell can be automatically detected, the color detection efficiency and quality of the electronic equipment shell are improved, and the cost is saved.

The application discloses a kind of LED light-emitting diode light color detection analysis method and system, it is related to LED detection technical field, including: light color acquisition module, for obtaining the spectral data generated in the lighting process of target LED;Timing modeling module, for in the continuous preset time period after target LED starts to emit light, multiple spectral data points are collected at fixed time intervals, main wavelength, color coordinates and fitting color temperature value are calculated to obtain, form multiple time light color parameter sequence, and construct time series model based on the sequence, output time series model and corresponding original parameter sequence.The application realizes the automatic judgment of the stable state of LED light emission by constructing the change curve of light intensity and color coordinates and introducing convergence trend analysis mechanism, effectively avoids the false determination caused by early data instability, and improves the representativeness and accuracy of detection data.

The present application discloses a method for detecting the quality of color wire harness wiring based on image recognition, which relates to the field of image technology. The method is based on the principle of image recognition, and utilizes a target detection model in combination with a semantic segmentation model to extract binary segmentation masks of connecting lines corresponding to various color labels. An initial color detection order is obtained through pixel-level detection. When it is determined that the initial color detection order does not match the predetermined color order, the color detection order is further adjusted and updated according to the position distribution of the binary segmentation masks of the connecting lines of each color label through pixel-level detection, so that the color detection order is consistent with the actual connecting line order, avoiding sequence detection errors caused by the cross-influence of different connecting line position distributions on image detection, thereby finally determining whether there is a wiring error. The method can quickly and accurately detect whether there is a wiring error and locate the connecting lines with wiring errors, with a high degree of automation and high detection efficiency.