Salah satu teknologi yang sedang berkembang pesat pada era modern revolusi 4.0 yakni kamera. Pixel (piksel) merupakan elemen terkecil yang diberikan oleh kamera citra digital yang bisa dilihat mata, piksel ini tersusun seperti matriks berisi angka 0 hingga 255. Piksel inilah yang dijadikan sumber data dalam proses pengolahan citra digital. Dalam penelitian ini dilakukan 2 proses pengambilan data yakni pengambilan data tinggi sebenarnya menggunakan alat ukur baku stature meter dan pengambilan foto sampel menggunakan kamera yang diletakkan pada jarak 160 cm dan tinggi 100cm dari sampel. Foto sampel yang diperoleh selanjutnya diproses untuk segmentasi/pemisahan tubuh sampel terhadap lingkungan sekitarnya menggunakan beberapa teknik pengolahan citra digital seperti grayscale, blur, deteksi tepi, dan bounding box sehingga didapatkan nilai piksel yang mewakili tinggi sampel. Tahapan selanjutnya ialah proses analisis regresi dengan menghubungkan nilai tinggi badan sebenarnya dengan nilai tinggi piksel dengan menggunakan 5 metode analisis regresi yakni metode persamaan kuadrat terkecil, metode persamaan berpangkat logaritmik, metode persamaan eksponensial, metode persamaan polinomial kuadrat, serta metode persamaan polinomial kubik. Penelitian ini melakukan analisis perbedaan nilai koefisien korelasi, Root Mean Squared Error (RMSE), error rata-rata dan akurasi yang diberikan antara kelima jenis analisis regresi tersebut dalam sistem perhitungan tinggi badan berbasis pengolahan citra digital terhadap nilai hasil pengukuran tinggi sebenarnya. Adapun hasil yang didapatkan ialah metode-metode ini memiliki perbedaan hasil nilai analisis yang tidak terlalu jauh/signifikan, metode berpangkat logaritmik dengan persamaan y=0.4825x^0.947 menghasilkan nilai analisis yang terbaik yakni nilai koefisien korelasi 0,.976, RMSE 1.3, error rata-rata 0.58%, dan akurasi 99.42%, Sedangkan metode polinomial kubik berada diposisi terakhir dengan nilai koefisien korelasi 0.978, RMSE 1.41, error rata-rata 0.64%, dan akurasi 99.36%.

One of the technologies that is developing rapidly in modern era of revolution 4.0 is the camera. Pixels are the smallest element provided by a digital image camera which can be seen by the eye, these pixels are arranged like a matrix containing the numbers 0 to 255. These pixels were used as data sources in digital image processing process. In this study, 2 data collection processes were carried out namely taking actual height data using a standard stature meter and taking sample photos using a camera placed at a distance of 160cm and 100cm high from the sample. The sample photos which obtained were processed for segmentation or separation of the sample body from the surrounding environment using several digital image of processing techniques such as grayscale, blur, edge detection, and bounding box so it can be obtained a pixel value that represents the sample height. The next stage was the regression analysis process by connecting the actual height value with the pixel height value by using 5 methods of regression analysis, namely the least squares, the logarithmic power, the exponential, the quadratic polynomial, and the cubic polynomial equation method. This study analyzed the difference in the value of the correlation coefficient, root mean squared error (RMSE), the average error and the accuracy given between the five types of regression analysis in the digital image processing-based height calculation system on the value of the actual height measurement results. The results obtained showed that these methods have differences in the result of the analysis values that are not to significant, the logarithmic power method with the equation y=0.4825x^0.947 produced the best analysis value, namely the correlation coefficient value of 0.976, RMSE 1.3, an average error was 0.58%, and an accuracy was 99.42%, while the cubic polynomial method was in the last position with a correlation coefficient value of 0.978, RMSE 1.41, an average error of 0.54%, and an accuracy of 99.36%.