Excessive bacterial load will not only cause delayed wound healing and local or systemic inflammatory reaction, but also threaten the life safety of patients in severe cases. Therefore, a new method that can detect wound bacteria quickly and directly is needed. Traditional bacterial detection methods are mainly visual observation of clinical symptoms (including fever, redness and swelling, pus exudate, etc.) and microbial swab sampling, but these two methods are usually subjective and time-consuming. The results of the same clinical symptom observed by doctors with different experiences may be different, and in some cases the infection may not show obvious symptoms. Therefore, it is often subjective to judge the infection by observing the clinical symptoms with naked eyes and may cause misdiagnosis. Swab sampling is often plagued by false negative, and different sampling methods and selection of sampling areas will affect the results of sampling and culture. False negative areas or missed areas may contain a large number of bacteria. The missing of a large number of bacteria in these areas may lead to repeated wound healing, causing great pain to patients and great burden on the medical and health system. In addition, longer culture time is also a pain point for swab sampling, because it may cause inaccurate culture results and miss the best treatment opportunity. Relevant studies have found that bacteria can emit fluorescence by themselves under the excitation of 405 nm light source, without using contrast agents. The purpose of this study is to design a fluorescence detection imaging device based on the principle of spontaneous fluorescence of bacteria, which provides a fast and direct new method for wound bacteria detection. In this study, a bacterial fluorescence imaging system is developed. The light source module of the system consists of two LED with a central wavelength of 405 nm and LED driver modules. After calibration and parameter setting, the light source driver works in constant current mode to ensure stable operation of the light source. When the light source shines on the wound bacteria, it can stimulate the spontaneous fluorescence of bacteria almost in a moment. The customized dual band filter is used to receive bacterial fluorescence and filter out the interference of reflected light from the light source. The transmission band of the filter is selected according to the results of the three-dimensional fluorescence spectrum of bacteria. In the dark environment, the smartphone imaging mode is set to night mode, the shutter speed is set to 3 s, the automatic white balance is set, and each sample is collected five times to reduce experimental error. Porphyrin fluorescence images are converted into three-dimensional intensity images to verify the imaging uniformity of the device. The fluorescence region is detected by region extraction algorithm, and the bacterial fluorescence is quantified. The signal area is extracted by using the Hough circle detection and edge contour extraction method as well as the mask method obtained by using the binary function. The gray value of all pixels in the signal area is traversed and the average value is calculated, which is taken as the quantitative result of the relative fluorescence intensity of bacteria. Finally, a linear rule between the fluorescence intensity and the concentration is obtained by linear fitting of the quantitative result. The experimental results of three-dimensional fluorescence spectrum of bacteria provide an important basis for the selection of excitation and emission bands in this study. It can be seen from the porphyrin three-dimensional intensity map that, except for the projection points of the two light sources, the uniformity of the signal area is good. This indicates that the two LED light sources used in the device can uniformly excite the fluorescent signal area, and the fluorescent image with good quality can be obtained without using the whole row of ring LEDs as light, which greatly reduces the power consumption and heat generation of the system. The result of image processing proves that regular and irregular signal regions can be extracted with the device. The regular region extraction algorithm can accurately extract the fluorescent signal region, avoid the fluorescent interference of the culture dish wall, and effectively improve the calculation accuracy of the relative signal strength. The gradient concentration of Escherichia coli and Staphylococcus aureus shows that there is a good linear relationship between the bacterial fluorescence intensity and the bacterial quantity under the experimental bacterial concentration. Combined with the bacterial plate counting experiment, the minimum detection limit of the system is 10⁵ CFU/mL, which indicates that the system can effectively detect the infection of wounds. Through linear fitting of the calculation results, it can be obtained that there is a linear relationship between bacterial fluorescence intensity and concentration, and the minimum detection limit of the system can be calculated according to the linear relationship. The device provides a new means for the bacterial detection of the wound, with high detection sensitivity to meet the bacterial load identification of the infected wound. There is no need for swab sampling and fluorescent labeling, and the detection procedure is simple and easy to operate. The extraction of signal region can clearly and intuitively show the location of bacteria, which can provide effective reference information for wound debridement and targeted drug administration. This study only conducts quantitative detection of bacteria in vitro, and further measurement of bacteria in vivo is needed. In addition, future efforts should include the development of a smartphone app that integrates existing processing algorithms to further simplify the process and enable rapid in vivo detection of wound bacteria.