Study On No-Reference Quality Assessment Method For Photoelectric Measurement Equipment Television Image

Photoelectric measurement equipment is an important component of the modernshooting ranges monitoring and control network, its main purpose is to obtain theflight path and pose of air targets, and characteristic of goals and other relatedinformation. CCD TV sensors is the main configuration of the photoelectricmeasurement equipment, the image is the final output, so the image quality is a keyindicator determining the overall performance of photoelectric measurementequipment. Over the years, image quality evaluation method of photoelectricmeasurement devices is not feasible, nor has the formation of evaluation criteria, so ithas not laws to the detection and acceptance of photoelectric measurement devices,which impacts the overall performance assessment to photoelectric measurementequipment seriously. Therefore, the research on the image quality assessment methodsof photoelectric measurement equipment, the exploration and establishmentassessment models and standards have important practical significance to step up thecapacity of national defense construction and modern shooting ranges monitoring andcontrol network.As in the actual photoelectric measurement equipment TV systems, the referenceor original images are not obtained basically, no-reference image quality assessmenthas practical meaningful to the television images. In the absence of any of the originalimage information, the construction of no-reference image quality assessment methods is much more difficult than full reference and reduced reference qualityassessments. In recent years, the research of no-reference image quality assessmentmethods has achieved some results, but these methods only for certain distortion typeshave not universality, which are independent of each other, and not form a recognizedresearch system.In this paper, the main factors arising image quality are analyzed, and the qualityassessment methods for photoelectric measurement equipment TV image areresearched thoroughly; the HVS-based model is introduced for the sake of combiningsubjective assessment with objective assessment effectively, three kinds ofno-reference quality assessments for TV image are designed and implemented fromthe three aspects (image contrast, sharpness and signal to noise ratio). In the imagecontrast assessment, an image contrast model based on HVS was established throughanalyzing contrast sensitivity function characteristic and local band-limited contrast, anovel no-reference image contrast assessment method under contrast sensitivity isintroduced. In the image sharpness assessment, an image sharpness model based onHVS is established through analyzing contrast sensitivity function characteristic andStructural Similarity, a novel no-reference image sharpness assessment method undercontrast sensitivity is introduced. In the image signal to noise ratio assessment, animage noise detection model is given through analyzing the visual masking, a novelno-reference image signal to noise ratio assessment method under visual masking isintroduced. Then a no-reference TV image quality assessment method is proposedintegrating the above three kinds of assessment methods. The assessment resultsobtained from the above four methods were proved to have good correlation with theobjective assessment results and in consistent with the human visual characteristics.