Research On The Luminance And Chroma Measurement Of LED Display Assembly Based On Color Image Sensors

LED display is an important media in the modern information society,and its application is becoming more and more widespread.The LED display is formed by the LED display assembly,and in the production process of the LED display assembly,the LED lamp bead has the luminance and chroma ununiformity.The problem becomes increasingly prominent as the LED display is high-definition and large-scale development.Therefore,it is necessary to measure the luminance and chroma of the LED display assembly efficiently and accurately.In this paper,based on the online detection of the LED display assembly,the luminance and chroma measurement model of LED display assembly based on color image sensors was studied.The luminance and chroma detection system of LED display assembly based on machine vision was designed and implemented.And the LED lamp bead batch classification problem was studied.The main research contents include:Based on the theory of photometry and colorimetry,the principle of LED display assembly and the imaging characteristics of color image sensors were analyzed.The luminance and chroma measurement model of LED display assembly based on color image sensors was constructed.The influence factors of the error in measurement model were studied through the simulation experiment.The results show that the optical system aberration,LED position error and image sensor noise have great influence on the error of luminance and chroma measurement of the LED lamp bead in different positions.Based on the simulation research,a visual inspection system for the luminance and chroma measurement of LED display assembly based on machine vision was designed.A dual camera exposure measurement scheme based on neutral gray filter was proposed,which overcomes the difference of luminance of LED lamp bead and the high dynamic range caused by light and dark switching detection.The motion control,display drive and image acquisition subsystems of the inspection system and the overall flow of the inspection system were designed.The complete algorithm of image processing of LED display assembly was designed,including image segmentation,positioning and gray level statistics.The collected images were analyzed to calculate the luminance and chroma reference values of all LEDs.The machines learning methods such as K nearest neighbor and support vector machine were introduced to study the multi-batch classification problem of LED lamp bead,and the classification accuracy rates of 99.95%and 99.54%were obtained respectively.Based oin the software engineering method,the luminance and chroma visual inspection system software of LED display assembly was designed.The software function and performance requirements were analyzed,the module was decomposed and designed,and each module was designed and coded in detail,and the performance and function tests were performed.According to the actual test results,the relative luminance measurement repeatability is 0.58%,the resolution is 0.34%,and the relative chroma measurement repeatability is 0.69%,the resolution is 0.33%,and the detection time of the single-chip LED display assembly is less than 6s.The system has a false detection rate of less than 4.7%for the luminance,chroma and mixed batch detection of the complete display assembly,and the missed detection rate is less than 2.2%.The system can meet the requirements of on-line detection of LED display assembly.